# Longeye --- Full Site Content > Longeye is an AI-powered investigative workspace that helps law enforcement agencies analyze digital evidence 100x faster. It processes video, audio, phone extractions, social media, financial records, and images, then returns verified, source-linked answers to natural language questions. Every finding is linked back to the exact moment in the original source evidence, making it easy to verify and defensible in court. Longeye is SOC 2 Type II certified, isolated per agency on dedicated US infrastructure, with no AI training on case data. --- ## Home Longeye turns mountains of digital evidence into actionable intelligence. Video, audio, phone extractions, social media, financial records: analyzed in minutes, not months. Every finding linked back to its original source. ### The Problem The justice system is fighting a new enemy: information overload. In a single case, investigators may be buried under 80+ hours of interviews, 10,000+ images (CSI photos), 250+ hours of jail calls, and 15,000+ hours of CCTV and body cam footage. Longeye processes it all, so your detectives don't have to. ### Capabilities Purpose-built AI that fits into investigative workflows. No training required, no complex prompts, no data shared. - **100x Faster Analysis** --- Longeye's AI-powered workspace rapidly digests and understands video, photography, and text at speeds up to 100x faster than a human reviewer. - **Every Evidence Type** --- One workspace for your entire case. Upload everything. Longeye processes it all and makes it searchable in minutes. - **Relevance Ranking** --- The most critical evidence is surfaced first, ranked by relevance to each specific case context. - **Verifiable Citations** --- Every answer links straight to the exact moment in a video, recording, message, or image. 100% verifiable and defensible. - **Language Support** --- Real-time language support translates and transcribes content across languages including Spanish, Mandarin, and more. - **Face Detection & Clustering** --- Automatically identifies and groups faces across thousands of photos and videos, surfacing every appearance of a person of interest. ### Trust AND Verify Every insight Longeye provides is 100% verifiable. Every piece of key intelligence is directly linked to the original sources so findings are easy to check and defend in court. - Direct links to original evidence sources - Relevance scores for every finding - Evidence-based report suggestions - Built for courtroom defensibility ### Proven Results Longeye is deployed across the country, from major metropolitan departments to state bureaus of investigation. - **$1M+ ROI Value Generated** --- A state department of corrections surpassed $1M in investigator time savings during their pilot alone. - **3 Staff Corruption Cases** --- Identified from a single query across 1,000+ jail calls. Cases that would have taken months to find manually. - **853 Hours of Audio Processed** --- A large Rust Belt police department fed 10,000 calls into Longeye. Evidence that would have never otherwise been found, surfaced in days. ### Wins from the Field - **South San Francisco PD (Homicide)** --- 25 hours of recordings reviewed in minutes. Detectives uploaded hours of jail calls and witness interviews. Longeye surfaced the key conversations and connections in a fraction of the time. - **Oklahoma DOC (Financial Crime)** --- Financial fraud uncovered across thousands of records. Longeye analyzed the data, identified suspicious transactions, and surfaced the evidence needed to build a prosecutable case. - **Oklahoma City PD (Homicide)** --- Confession found buried in a no-DNA case. A single query surfaced a recorded confession buried deep in the evidence, turning a cold investigation into a closed case. ### Why Longeye vs Civilian AI Tools **Longeye:** - No complex prompts needed, just ask a question - Every finding linked to source evidence - Isolated per agency. Your evidence never commingled - Video, audio, images, documents & social media - 100+ language translation - Case-level AI chat across all files - SOC 2 Type II certified - No data used for AI training **Civilian AI Tools:** - Requires expert-level prompting for results - No source citations or verification - Data sent to third-party servers - Limited to text or single-file analysis - Not built for law enforcement security standards - Your data may train their models ### Security Security built from the ground up for the most sensitive evidence. Designed from the ground up with defense in depth: Live Scan Staff, SOC 2 Type II, Private AI, Hosted on AWS, Isolated per Agency, No Data Training. ### Testimonial "Longeye has already proven its value, cutting months off of investigations, surfacing leads we might have missed, and even closing gaps in a cold-case homicide. Every hour Longeye gives back to my detectives is an hour invested in solving cases, supporting victims, providing closure and strengthening community trust." --- Chief Darrell Lowe, Redmond Police Department --- ## Product Ask any question about your case. One prompt. Every evidence type. Longeye searches across your entire case: video, audio, phone extractions, financial records, social media. And returns verified, source-linked answers. ### Features - **Rapid Evidence Ingestion** --- Drag and drop thousands of files at once. Longeye digests video, photography, audio, social media, cell phone data, and messages in seconds. - **Contextual Search** --- Ask questions in natural language. Longeye contextually reads evidence to find key intelligence, ranking every finding by relevance. - **Source-Linked Citations** --- Every answer links to the exact moment in a video, recording, message, or image. Findings are easy to check and defend in any legal proceeding. - **Real-Time Translation** --- Translates and transcribes content across languages including Spanish, Mandarin, and more. - **Automated Report Writing** --- Summarizes key evidence and helps detectives quickly write case reports, suggesting evidence-based additions. - **Isolated Investigations** --- Privacy by design ensures AI models never train on case data and every investigation remains fully isolated. ### Investigative AI Type a question in plain language and Longeye searches every video, recording, message, and image in the case. Findings are ranked by relevance and linked directly to the exact moment in the source evidence. ### No Prompting Expertise Required You shouldn't have to memorize a three-paragraph prompt to get good information out of your evidence. Longeye is designed so any investigator can get powerful results from simple, natural questions. ### Audio & Media Analysis Longeye transcribes and translates audio and video across languages, then surfaces every call on an interactive timeline, color-coded by AI relevancy score. Searchable transcripts across Spanish, Mandarin, and more. Call history timeline color-coded red/yellow by relevancy. ### Geospatial Intelligence Longeye plots location evidence on an interactive map: GPS tracker pings, Cellebrite extraction data, CDR lat/long, and ALPR hits from networks like Flock Safety. A custom timeline slider animates movement over time across up to 10,000 data points with direction-of-travel arrows and automatic reverse geocoding. ### Financial & Tabular Analysis Upload any spreadsheet or CSV (bank records, phone tolls, inmate commissary logs) and Longeye analyzes it natively. Ask a question in plain English and get an AI-generated chart, table, or summary back in seconds. ### How It Works Instant onboarding. Multiple agencies have closed real cases in their first hour using Longeye. No integrations required. No training needed. Just upload evidence and start asking questions. --- ## Investigate Longeye Investigate is the case-centric product for law enforcement, at https://www.longeye.com/investigate. A detective drops the whole case into one workspace, phone extractions, jail calls, video, social media warrant returns, and records, and Investigate sorts it into Explorers, scores every file from 1 to 10 with a reason, and answers plain-language questions with source-linked citations. Capabilities include relevancy scoring, thematic search, geospatial mapping, jail-call review at scale, and search-warrant drafting where every claim links back to the evidence. --- ## Monitor Longeye Monitor is the proactive product for corrections intelligence units, at https://www.longeye.com/monitor. Instead of waiting for a tip, Monitor reads every call, message, and visit a facility produces and surfaces the activity worth watching before it becomes an incident, ranked by relevance with the reason on every flag. An investigator can open a case straight from a flagged conversation, and Monitor connects calls, messages, and visitor patterns in one question, each finding linked back to the original record. --- ## Law Enforcement The law enforcement page, at https://www.longeye.com/law-enforcement, positions Longeye Investigate for city and state detectives, investigators, and analysts facing the Big Evidence problem. Drop in the whole phone dump, the jail calls, the warrant returns, and the video, and the evidence that matters comes to the top, scored and explained, with every answer one click from the source. --- ## Prosecutors The prosecutors page, at https://www.longeye.com/prosecutors, positions Longeye Investigate for District Attorneys, Prosecuting Attorneys, and Attorneys General. It is organized around the case lifecycle, charging, discovery, and trial, so the decision rests on the whole record, with every finding linked to the source for verification. --- ## Public Defenders The public defenders page, at https://www.longeye.com/public-defenders, frames Longeye Investigate for the defense. The discovery is just as large for a public defender, with a fraction of the time, so Longeye reads the whole file the state turns over and helps the defense find what matters for the client and check the state's account against the source. The Public Defense Grant covers a full year of Longeye for the defense. The application page at https://www.longeye.com/public-defense-grant collects an office's details and routes them to the public-defense team. Eligible applicants include public defender offices at the county, state, and federal level, indigent-defense and exoneration nonprofits such as the Innocence Project, and the criminal defense investigators who work for those organizations. Approved offices receive a $50,000 grant covering a full year of the full Longeye platform. --- ## About At Longeye, we're taking a stand for truth. ### Our Mission We think like detectives. And build like engineers. Longeye addresses a critical challenge facing modern law enforcement: the overwhelming volume of digital evidence in criminal cases. We believe that the certainty of being caught is one of the greatest deterrents to crime. ### Our Values - **Balanced Justice** --- We're committed to strengthening the integrity of the justice process for all. Longeye plans to provide AI services free of charge to public defenders' offices. - **Built With Investigators** --- A team that listens to the people we create products for. Longeye is the product of working side-by-side with police on the ground. - **Privacy First** --- Our AI models never train on case data and every investigation remains fully isolated. Security built from the ground up with defense in depth. ### Team - **Guillaume Delepine, Cofounder & CEO** --- A public safety and social justice advocate with a decade at the forefront of technology. Previously spent six years at Skydio building the public safety business. Princeton alumnus. - **Danielle Man, Cofounder & CTO** --- A seasoned technologist. Before co-founding Longeye, she was Senior Director at Apollo GraphQL where she led the 40-person engineering group. - **Matt Hoffman, Head of AI** --- Leading AI developer with experience at Weights & Biases, Snorkel, and Primer AI where he rose to Director of Professional Services. Cornell & Duke alumnus. ### Investors Longeye has raised $5M in seed funding led by Andreessen Horowitz's American Dynamism Fund, with additional investment from Seven Stars Capital and strategic angels. Founded in 2025. --- ## Careers Longeye is hiring. The careers page at https://www.longeye.com/careers covers what we build, what customers say about it, how agencies grow with us, how the team works, press coverage, and the people behind the product. We're a small team in San Francisco: engineers, go-to-market people who spend their weeks inside police departments, and a couple of former operators who used to carry a badge. Customer feedback goes straight to the person who built the thing, everyone sees the same numbers leadership sees, and output matters more than hours or location. Open roles are posted on our job board at https://ats.rippling.com/longeye/jobs. --- ## Contact Ready to see Longeye in action? Reach out for a demo or to learn more. - **Headquarters:** San Francisco, California - **Email:** info@longeye.com Whether you're a law enforcement agency, a public defender's office, or a technology partner, we'd love to hear from you. As part of our commitment to balanced justice, Longeye plans to provide AI services free of charge to public defenders' offices. --- ## Live Walkthrough Law enforcement agencies can sign up for a live group walkthrough of Longeye with our team at https://www.longeye.com/walkthrough. In a 30-minute session, our team runs the platform on investigative evidence and answers workflow-specific questions. Agencies that attend are offered a two-week on-site trial at no cost, with our team training investigators in person and leaving Longeye running on active cases. --- ## IACP Technology Conference Longeye exhibits at the IACP Technology Conference at Booth 629. The dedicated landing page at https://www.longeye.com/iacp lets public-safety tech leaders book a 1:1 meeting at the booth or join a virtual walkthrough. Longeye is positioned as the secure alternative to civilian AI tools like ChatGPT and NotebookLM: isolated per agency on dedicated US infrastructure, SOC 2 Type II certified, no data used for AI training, and every finding linked back to source evidence. The page targets CIOs, IT directors, and command staff who need an AI platform their legal, IT, and command staff can all sign off on. --- ## FBI National Academy Session 298 Longeye supports FBI National Academy Session 298 with a dedicated landing page at https://www.longeye.com/fbinaa. NA session attendees can sign up for a 30-minute virtual Longeye session run live by the team, with Q&A tailored to their agency's workflow. Sessions are followed by a two-week on-site trial offer at no cost. The page reinforces Longeye's positioning for public-safety leaders: secure AI that legal, IT, and command can all sign off on. --- ## For Corrections Longeye for corrections at https://www.longeye.com/corrections is the dedicated landing page for state DOCs and county jail intelligence units. The page reframes the volume problem inside the walls: corrections isn't a one-case-at-a-time job, it's a continuous stream of inmate calls, tablet messages, and visitor logs across an entire facility, and most of it never gets reviewed. Longeye reads every recording, message, and visitor record the facility already produces, transcribes and translates across 100+ languages including coded language and slang, and links every finding back to the exact second of audio or line of a message. The page documents real outcomes from a state corrections department in active pilot: 853 hours of inmate audio reviewed in a single sweep, 15,000 hours of detective-labor equivalent processed in six weeks, $4M projected annual workforce productivity. Case studies include a new-hire contraband-smuggling plot caught before her first day, real-time apprehension support during three active prison escapes over eight days, and a covert correctional-nurse/inmate relationship uncovered after only two of eight call batches. Plugs into existing call-recording exports today, with more integrations coming. Isolated per agency, SOC 2 Type II, federated authentication via Microsoft Entra ID. Form submissions route to the corrections-specific Brevo segment for outbound sequencing. --- ## Security The security page at https://www.longeye.com/security is the single public reference for how Longeye protects evidence, written for the IT, legal, and command staff who sign off before an agency adopts AI. It lays out the foundations: SOC 2 Type II, isolated per agency on dedicated US infrastructure, encrypted at rest and in transit, SSO with MFA and role-based access, a 1-year immutable audit trail, AI that never trains on case data, US data residency, background-screened staff, and resilient WORM backups. A dedicated section covers verifiable AI and the patent-pending citation engine that grounds every answer in cited source evidence, so a finding can be opened, checked, and reproduced instead of trusted on faith. The page then breaks out what law enforcement, prosecutors, public defenders, and corrections each need to know, and points agencies entering a trial or already under contract to the full trust center, which they can request through the security request form at https://www.longeye.com/security-request. --- ## Blog Posts ### Funding Investigative Technology with FY26 COPS TEP (2026-06-12) The COPS Office [FY26 Technology and Equipment Program (TEP)](https://cops.usdoj.gov/tep) directs funding to law enforcement agencies for the technology and equipment that help them respond to and prevent crime. For agencies with a technology or investigative project title, that funding can go toward one of the most stubborn problems in modern casework: the sheer volume of digital evidence. ## The Problem Funded Technology Should Solve We have entered an era of Big Evidence. A single case can carry hundreds of hours of jail calls, phone extractions that run tens of thousands of pages, and warrant returns no human can read end to end. Most of it never gets looked at, not because investigators do not care, but because there are only so many hours in a shift. A detective opens the files that load fastest instead of the ones that matter most. That is the work funded technology should attack, because it is the work that quietly stalls cases. ## What Your TEP Award Can Fund Software is an [allowable cost](https://cops.usdoj.gov/pdf/2026ProgramDocs/tep/Resource_Guide.pdf) under TEP. Evidence-analysis tools map cleanly to project titles like technology upgrades, investigative technology, software upgrades, and digital evidence. So if your agency holds one of those, an analysis platform is squarely inside what the funding is meant for. Longeye is one option built for exactly this. Investigators drop in the whole case, calls, messages, photos, video, documents, and device extractions, and ask plain-language questions. Longeye does the grunt work of going through all of it and answers with source-linked findings. An Oklahoma City detective cut jail-call review from more than 20 hours a week to under five. ## How to Structure It One rule shapes how software should be purchased here. A multi-year or subscription contract is allowable only if the full term is paid inside the 24-month award period. So the clean way to buy is a fixed, two-year prepaid term, billed once and paid within the window. A two-year software line item usually stays under the $350,000 threshold that triggers [sole-source review](https://cops.usdoj.gov/pdf/2026ProgramDocs/Sole_Source_Justification_Fact_Sheet.pdf), which keeps the path simple.
A note on the fine print: TEP does not cover indirect costs or costs incurred before the award start date, and funds cannot replace money the agency was already going to spend. Price the purchase as a direct, net-new cost inside the award window.
## Built to Stay on the Allowable Side of AI Agencies are right to scrutinize AI, and TEP draws a clear line: facial recognition is unallowable, and biometric technology is rarely approved. Evidence analysis sits on the other side of that line. Longeye only works through evidence an agency already lawfully holds, and every finding is one click from its source, because you cannot put an algorithm on the witness stand. It is SOC 2 Type II, self-hosted, and never trains on case data. ## What to Do Before July 23 Two practical steps. First, tie the purchase to your specific project title, the scope your agency was [designated for in the FY26 Joint Explanatory Statement](https://www.appropriations.senate.gov/imo/media/doc/fy26_cjs_jes.pdf), so the budget line maps to what the award was meant for. Second, make sure both of your Authorized Representatives have logged into JustGrants and that your SAM.gov registration is active, since that is a common last-minute blocker. The Grants.gov first step is due July 16, the final application July 23, and awards are announced on or after October 1. --- Longeye is an AI evidence-analysis platform for law enforcement, used by agencies across the country and covered by the Washington Post, KING 5, and ABC7. Software is an allowable TEP cost, and we structure terms to be paid in full within the award period.

Holding a TEP award with a technology title?

See Longeye run on the kind of case your unit works, then we will help you fit it to your project scope before July 23.

Request a demo
--- ### Can Investigators Use AI Without Breaking CJIS? (2026-06-05) The most recent version of the FBI's CJIS (Criminal Justice Information Services) Security Policy, version 6.0.0, was published in December 2024. By that point, ChatGPT had been generally available for over two years. Yet "artificial intelligence" appears in the policy exactly once, and only in the context of using AI for malicious code detection.
The federal rulebook for handling criminal justice information mentions AI exactly one time. That single mention is about using AI to detect malicious code, not about detectives using AI on case files. The policy that governs CJI handling has not yet caught up to how investigators are actually using these tools today.
In the time since, many detectives have started turning to AI to help work through large amounts of digital evidence. Commercial AI tools designed specifically for investigations, like Longeye, have only recently become available. In the meantime, some investigators have reached for general-purpose products like ChatGPT, Claude, Meta AI, and Grok. The obvious problem with that kind of early adoption is that the information detectives are feeding into public products is, by definition, criminal justice information. CJI is not classified the way military intelligence is, but it is extremely confidential. Think of it as sitting somewhere above credit card data (PCI) and personal health information (HIPAA), and below national security material. A disclosure can put witnesses and confidential informants at risk. Even the fact that an investigation exists can compromise it, prompting suspects to change their behavior or destroy evidence. CJIS is a demanding standard that criminal justice agencies are required to follow, often without the staff or budget to follow it well. Smaller departments simply cannot afford a full-time compliance officer and the IT security posture that compliance requires. Compliance work can take a back seat to the need to close cases. The risk isn't theoretical. Failing a CJIS audit can cost an agency access to national databases like NCIC (the National Crime Information Center). Officers who disclose CJI can face criminal and civil penalties. It helps to remember that detectives are information workers, with the notable exception that they show up to work armed. They use computers. They get locked out of their accounts, including the FBI director on occasion. They organize files, search the web, and order from DoorDash. It is natural for a detective to be curious, even enthusiastic, about using AI to draft a police report, search a 100,000-page warrant return, work through a cell tower dump, transcribe an interview, or do any of the other tasks that consume their week. The security and accuracy stakes, though, are much higher than in typical office work. ## Basic Security for CJI Any system that processes CJI, AI or not, has to secure it from unauthorized access. CJIS spends hundreds of pages on the details: authentication, access control, least privilege, encryption at rest and in transit, key management, firewalls, intrusion detection, incident response, and everything else the FBI could think of in 2024. These are table stakes for any software provider that works with criminal justice agencies. ## AI Concerns Not Directly Covered by CJIS Across the criminal justice agencies we work with, two AI-specific concerns come up again and again. The first is the security of CJI inside the model itself, meaning who has access to the AI's entry and exit points and to its internal state. The second is the risk that CJI could end up in training data and become embedded in future public models. Exposing public AI products to CJI practically guarantees that the information is disclosed to unauthorized personnel at the provider, and that it can leak into future training sets. Even the prompts a detective enters can themselves be CJI.
A prompt like "Find instances in the evidence where suspect [name] encountered victim [name] before [incident]" is itself CJI. The names, the relationship, and the fact that an investigation is open are all sensitive. Type that into a public AI tool and you have disclosed an active investigation to a third party.
To avoid all of this, Longeye runs AI models in a secure environment where model providers have no visibility into the data or the prompts, and where no data can flow back to the provider. We use Amazon Bedrock to run commercial and open-source models. We also run specialized models on Amazon-hosted GPUs for cases like speech-to-text and image analysis. Customer data belongs to the customer. It is not used to train or update models, and there is no path for it to flow back to the model providers for training under any circumstance. ## Hallucinations The current CJIS policy was written before hallucinations became a recognized concern in AI. The horror stories are easy to find. Lawyers who used AI to prepare briefs and discovered, in front of a very angry judge, that the citations were fabricated. It has happened in high-profile cases brought by the Department of Justice itself.
An imagined fact in a police report is not a typo. It can wreck a case at trial, and if the defense never catches it, it can contribute to a miscarriage of justice. The AI tool an investigator chooses has to be incapable of inventing evidence.
An investigator cannot afford imagined facts in a report. A hallucination could wreck a case at trial or, if the defense never catches it, contribute to a miscarriage of justice. Longeye protects against this by always grounding its answers in the original evidence. When the system surfaces a relevant detail, summarizes a document, or helps draft a report, it cites the source it came from. If a claim cannot be traced back to a real file, Longeye will not make it.

Evaluating AI for your agency?

If CJIS posture, training-data exposure, and hallucination risk are on your checklist, we can walk you through how Longeye handles each one and sign a CJIS Security Addendum at onboarding.

Talk to us
--- ### June Product Update (2026-06-04) ## Relationship graphs that link back to the evidence Chat has a new trick: it can draw the relationship graph for you. Picture a case with a few hundred phone numbers and no obvious shape. Ask Chat who's connected to whom, and instead of a wall of digits you get a picture: who's talking to whom, and who sits at the center of it all. The phone that everyone routes through tends to be the one you want. Here's the part we're proud of. As Longeye processes a case, it's not just noting that two numbers talked. It's gathering what it can about the people behind them. So the graph isn't a bare web of dots. Click a node and you get a sense of who that person is and what the relationship actually is, not just that one exists. And every link is source-linked: click a connection and you land on the evidence that drew it, whether that's the call, the PDF page, or the message. A graph won't even draw a link it can't point back to. Zoom controls keep dense networks readable. ## A home for your reports: the Drafts page You've been able to write reports in Longeye for a while. This month they got a front door. The new **Drafts** page gathers every report for a case in one place, so you're not hunting for the one you started last week. Starting a new one is a click: open Drafts, hit **Create report**, give it a title, and you're writing. You can pull in specific files right from the create dialog for Chat to draw on as you go. ![The Drafts report editor: an affidavit reviewed section by section, with accepted and pending-review labels, a section-progress bar, and an edit timeline showing who changed what](/images/blog-jun-drafts-review.webp) The bigger change is how a report gets finished. A report comes broken into sections, and each one is reviewed on its own: accept it, send it back, or leave it pending. The bar across the top tracks how far along you are: two accepted, three to go. Every edit is recorded with the section, the version, and who made it, so a partner can pick up where you left off and you can see exactly how each line got there. That trail isn't just convenient. It's what an AI-assisted report needs to hold up later. ## CSAM detection and reporting Longeye has flagged suspected CSAM automatically during processing for a while now. This month we focused on two things: making the detection itself more precise, and making the human review faster and the reporting automatic. **Sharper detection.** Early on, detection erred toward over-flagging, and the false positives added up. We narrowed the criteria and added a second-stage visual verification before anything gets flagged, so the flags you see now are far more likely to be real. This came directly from your feedback. **Review in place.** Suspected CSAM appears behind a blurred, hazard-striped frame everywhere it could surface (gallery, lightbox, folder view), so it's never shown clearly by accident. You confirm or dismiss a flag without leaving the image explorer, and the blur preference stays in sync across views. **Confirm once, propagate everywhere.** When you confirm a flag, that file's hash is recorded, and any other file matching it, across every case in your agency, is flagged automatically. You don't re-review the same image twice. **Automatic NCMEC reporting.** Confirming a file files a report to NCMEC's CyberTipline on your behalf, with a full audit trail of who confirmed it and when. The mandated report goes out without a separate manual step. ## More Chat improvements - **Batch image description.** Ask Chat to describe a set of photos and get one summary table back, up to 25 images in a single turn. - **Search by capture date.** "Videos from September 20" or "calls on the 14th" now works on every case, including ones full of hash-named forensic files. - **Timezone conversion.** Ask Chat to convert a time to any timezone and it computes the answer instead of describing it in prose. ## Other improvements - **Key moments on video.** The video scrubber now marks the most investigatively significant moments. Hover to preview, click to jump. - **Per-channel DVR.** Multi-channel DVR recordings split automatically into one video per camera channel, each with its own thumbnail, transcript, and analysis. - **Filter calls by caller PIN.** On Securus and Viapath jail-call cases, filter the audio explorer by caller PIN to isolate one inmate's calls. - **PDF viewer redesign.** The document detail page now has Chat, Overview, and Themes tabs with a page minimap for moving through long reports. - **Apple Mail support.** .emlx files from iOS backups and UFDR exports now show up in the Documents explorer, ready to search and chat about. - **SHA-256 on every file.** Each file's detail page shows its SHA-256 hash with a copy button, for chain-of-custody records. - **Recover interrupted uploads.** If a session drops mid-upload, a recovery page shows exactly which files didn't make it through, so you re-upload only those. ## Under the hood Plenty more landed this month that doesn't need its own headline. Chat got steadier across the board: fewer dead ends on tricky questions, citations that point where they should, and answers that hold up on long and unusual files. Search and playback got more reliable, and a good stack of smaller rough edges across audio, images, documents, and chat got sanded down. A handful of quieter wins worth calling out: Chat carries recent results across turns, so follow-ups don't re-run the same search; calls that switch languages mid-recording translate in full; image search reads meaning, not just exact words; and your filters now survive clicking into a result and back. Large audio cases got faster, too: UFDR archives with thousands of jail calls now spread across parallel workers, so a big extraction clears in a fraction of the time. --- As always, find us at support@longeye.com if you have questions or feedback. --- ### May Product Update (2026-05-06) In May, the days are longer and the evidence queues are too. This month's update is built for that moment: faster video review, steadier performance on large cases, and broader support for the messy file formats investigators actually see in the field. ## Video, rebuilt The big thing this month is that our support for Video now **watches** your videos, not just **listens** to them. That means you can get through piles of surveillance footage in hours instead of weekends. One investigator used it on bus footage and, within minutes of logging into Longeye, found the exact moment their person of interest boarded the bus. ![Video detail page showing extracted scenes beside synchronized playback and a cited Chat answer](/images/blog-may-video-scene-extraction.png) The video list is now a gallery grid, with thumbnails and auto-generated scene titles. Rather than a flat list of filenames, you see what's actually in each clip. Theater mode provides full-screen playback with a karaoke-style synchronized transcript — so you can watch and read at the same time without switching between views. Chat now analyzes video scene by scene, using a dedicated GPU pipeline we've been building for months. That means the heavy visual AI work runs on infrastructure built for this kind of footage review, so Longeye can search what is happening on screen instead of relying only on transcripts. Ask what's happening at a specific point in the footage, and you get a cited summary with timestamps that jump directly to that moment. Every citation is clickable. Picture-in-picture lets you keep a video playing while you work elsewhere in the case. The video and transcript panes are resizable. Playback speed goes up to 16x. Exotic formats — VOB, MKV, AVI, WMV, and Dahua's .dav format from commercial security systems — now process automatically. ## Chat Chat can now extract IP addresses from any document, spreadsheet, or database and plot them on a map. Useful for login records, email headers, and connection logs. Citation reliability got a significant upgrade. Chat now keeps tighter track of the source material behind its answers, making document, call-record, table, and video citations easier to inspect and less likely to point at the wrong evidence. Chat also handles bigger investigative questions more gracefully. Warrant drafts, CDR cross-referencing, and broad document searches can now work through more sources and larger tool results before summarizing, instead of stopping early when a case has a lot of data. The chat session history sidebar has a new layout: sessions organized into Shared with Team, Starred, and Recent, with small icons showing when a conversation produced a map, chart, or table. ## New file and evidence types A lot of new formats ship with this update: - Viapath jail call records, with full metadata: caller name, PIN, recipient, facility, and call time - ALEAPP exports (Android forensics tool widely used to parse phone logs), which were previously dropped silently at upload - Instagram DM exports - Google LERS search warrant returns, including email record sets - Password-protected ZIP files, so locked evidence archives can be unlocked and processed without unpacking them locally - PowerPoint (.pptx) and Word template (.dotx) files - Dahua security camera footage (.dav) - MPEG Transport Stream (.ts) video Audio now gets an automatic type classification — interview, body cam, 911 call, and more — with a matching filter on the audio page. AXIOM portable cases now show cleaner file paths, human-readable device names, and extraction metadata including BFU/AFU lock state, extraction scope, and chip family. ## Under the hood We also kept pushing on large-case performance. The images page now loads much faster on very large cases, including the complicated filter combinations investigators use to narrow hundreds of thousands of images without waiting on every count to recalculate. Other fixes this month: random session logouts during long video reviews, folder filters returning zero results on large cases, CDR map timestamps now auto-detecting local timezone from GPS data, and upload flow now shows actionable error messages, including a warning when a partially-downloaded file is uploaded by mistake. --- ### The Washington Post: Detectives test out a potential crime-fighting partner: AI (2026-04-10) The Washington Post published an in-depth look at how law enforcement agencies across the country are beginning to use AI tools like Longeye — and the thorny legal questions those tools raise in courtrooms. Reporter Katie Mettler profiles Detective Lauren Cunningham of the Oklahoma City Police Department, who initially approached Longeye with skepticism. The 20-plus hours per week she would normally spend monitoring jail calls from murder suspects had been reduced to less than five. A sex crimes investigator used Longeye to translate 10 suspect phone calls, uncovering a confession that turned a child rape case headed for trial into a likely plea agreement. The story places Longeye in a broader landscape of AI tools being marketed to law enforcement — license plate readers, facial recognition, ballistics analysis — many of which have drawn scrutiny from civil liberties advocates. Longeye's approach, built on a closed sandbox using only evidence investigators upload, is presented as a deliberate response to those concerns. CEO and founder Guillaume Delépine tells the Post: "Dirty doesn't work in the justice system. You have to build a much more deeply thinking machine." The article also covers the Oklahoma Department of Corrections, which is using Longeye to sift through the nearly 50 million minutes of prison phone calls it monitors annually, and notes that in 2026 Longeye has streamlined approximately 34 years' worth of detective work into just a few months — processing 25 million files across 35 law enforcement agencies. [Read the full article in The Washington Post](https://www.washingtonpost.com/national-security/2026/04/10/ai-police-criminal-investigations/) --- ### April Product Update (2026-04-08) Spring brought a large wave of new capability to Longeye. Here's what we shipped in March. ## Tables: Structured Data Analysis with AI ✨ This month we launched Tables, a dedicated workspace for tabular data. We know that every case has an enormous amount of this data, whether it's activity data from a phone, messages from Snapchat, or records from a GPS tracker. ![Tables view showing GPS tracker data with an interactive map visualization](/images/blog-apr-tables.png) Now you can chat directly with your tables — add columns together with *real math*, plot trends on charts, or even render routes on a map. **What's included in the Tables release:** - **Cross-file queries.** Chat can query across every table in a case simultaneously, not just one file at a time. This makes it practical to work with phone extractions where data spans dozens of related spreadsheets. - **Court-ready evidence packets.** We know your evidence needs to be admissible in court. Every time Longeye creates a map or plots a chart, that artifact comes with a downloadable evidence packet that walks you through the exact steps to reproduce it outside of Longeye — so you know exactly how it was made and can make sure it was done right. - **GPX file support.** GPS track files from phones and wearables now parse into the Tables view automatically, enabling map-based and time-based analysis alongside other structured data. - **Date range display.** Each table shows the temporal span of its data — the earliest and latest dates in any time column — so you immediately know whether a file covers the relevant time period without opening it. - **System table suppression.** Phone extractions include many internal system files with no evidentiary value. These are now hidden by default, reducing noise by roughly 86%. - **Persistent table and chart history.** Tables and charts generated in Chat are saved in the conversation history and reload when you return to a session. ## Images 🖼️ Longeye now extracts text from images automatically. If the text is in another language, it's translated to English. All extracted and translated text is fully searchable, so you can find images by what's written in them — regardless of language. ![Extracted and translated text from a Chinese iMessage conversation](/images/blog-apr-images.png) We also made a number of other improvements to images this month: - **Date filter.** A new date filter on the Images page lets you focus on photos with or without a capture timestamp, or narrow to a specific date range. Dates are displayed in the case timezone with an indicator when the source timezone is unknown. - **Smarter relevance scoring.** Longeye now factors in where a photo lived on the device before scoring its relevance. A photo from the camera roll is treated very differently from one found in an emoji cache or app icon folder. We measured, and it showed a 78x difference in relevance rates based on these signals — you should see noticeably less noise in image results. - **Scroll position indicator.** A floating position ticker appears when scrolling through large image sets, showing the current position and offering a "Jump to…" control to skip to any image directly. ## Audio and Transcripts We're continuing to invest in our bread and butter: audio analysis. You can now edit transcripts and save those edits for your teammates. You can also, finally, give names to the speakers in audio transcripts — so you don't just see "Speaker 0" and "Speaker 1." A dropdown shows all existing speakers in a recording, and clicking one renames all their segments in a single step. We also improved transcript search — you're less likely to miss results that fall near the boundary of a long recording. ## Case Management **Case type.** Cases can now be categorized by investigation type — Homicide, DV/Stalking, Fraud, CSA, and nine other categories. The type shows as a badge in the cases list and is filterable. **Case lead.** Cases now have an assignable lead, separate from whoever created the case. When a case gets handed off — a detective rotates out, a new analyst picks it up — you can reassign the lead so it shows up in the right person's "My cases" tab. **Duplicate file detection.** Longeye now automatically detects duplicate files across your uploads. When a file has already been processed, we skip reprocessing it entirely — saving you both processing time and usage. Duplicates are flagged on the Evidence page so you always know what's original and what's a copy. ## Under the Hood As always, a large part of the month went into performance and reliability. The Images page loads substantially faster on large cases — an 83K-image case that previously took 24 seconds now loads in a fraction of that time. We fixed source path filters that were showing empty results despite correct counts in the sidebar, resolved audio page loading spinners that got stuck when combining filters with deep links, and squashed upload failures caused by filename collisions across nested folders. PDF and large phone extraction processing is more resilient, with fewer timeouts on complex files. We love hearing how investigators are using these features in the field — if something saved you time on a case or you have an idea for what we should build next, drop us a line. And if you haven't tried Longeye yet, [let's set up a demo](/contact) and we'll walk you through it. --- ### ABC7 News: How Longeye's AI is Helping Law Enforcement Crack Cases Faster (2026-03-28) ABC7 News reports on how Longeye, a Bay Area AI startup co-founded by Guillaume Delepine, is transforming how law enforcement processes digital evidence. The segment features a live demonstration with South San Francisco Police Captain Anthony Penel, showing how Longeye analyzed 100 jailhouse phone calls from a felony case in seconds, identifying a critical conversation where suspects coordinated an alibi in Spanish. The story highlights how 70 percent of investigators today do not have enough time to review all of their digital evidence. With surveillance footage, phone calls, texts, emails, and massive PDFs, the volume of potential evidence has exploded while detective staffing has stayed flat. Longeye streamlines this process, connecting dots across thousands of pages and hours of recordings that would take human reviewers days or weeks. Newly retired San Ramon Police Chief Denton Carlson, who joined the Longeye team, calls the technology game-changing for law enforcement. The segment also highlights a case where a department's interpreter missed a confession buried in 30 hours of recorded calls that Longeye flagged immediately. ## Full transcript Good old-fashioned police work solves crime. But in an increasingly digital world, shoe leather isn't always enough. Enter Longeye, a Bay Area AI startup co-founded by Guillaume Delepine, using artificial intelligence to help street-smart cops catch crooks. **Tell me about the concept.** Longeye is based on the fact that 70% of investigators don't have time to get through all of the digital evidence that they have to go through. There's just too much potential evidence to examine. On any given case, we're talking about surveillance camera footage, hours of phone calls, thousands of texts and emails, more evidence than ever. And in most departments, no more detectives than years ago. When newly retired San Ramon Police Chief Denton Carlson saw the technology, he wanted to be a part of it and joined the Longeye team. The technology, he believes, is game-changing for law enforcement. **You're one click away and you're going to the actual evidence to verify exactly what the AI has found for you. Not just searching the random internet. You're going to the evidence that you have inputted, basically.** Exactly, with clear citations and direct links. Captain Anthony Penel of South San Francisco Police was part of a real-world Longeye demonstration, using the AI tool to comb through 100 jailhouse phone calls from a suspect in a felony case. **The system was able to point directly to the most relevant call, which ended up being one of our suspects discussing their alibi through a third party to plan an alibi with one of the other suspects.** So what would have taken detectives hours and hours, if not weeks, to sort through, the Longeye AI tool did in just seconds. And what's more, it found that conversation in Spanish. It can do this in 100 languages. Longeye co-founder Guillaume Delepine says the sheer amount of electronic evidence today is overwhelming. His creation streamlines the process and connects the dots. **That sometimes are in 30,000-page PDFs, thousands of pages apart, all brought into one bullet point that helps the detective understand what's going on on the ground.** Giving law enforcement a powerful new tool that changes how police do their jobs and how the bad guys get caught. **You're making it very difficult for me to pursue the life of crime I'd always imagined.** Here's another example. One department hired an interpreter to go through 30 hours of recorded phone calls. The interpreter found nothing. When they ran it through the Longeye platform, it immediately flagged a confession from the suspect that the interpreter had missed. Longeye continues to expand to more than a dozen police departments around California and Arizona, with many more on the horizon. --- ### March Product Update (2026-03-01) Spring is in the air, and so are new features! Happy March, everyone. ## Call history timeline When you review large volumes of jail calls, Longeye now plots repeated calls to the same phone number on a timeline. We highlight relevant calls, and you can click different dots to jump from call to call quickly. We're hearing this has been really helpful for local PD's collaborating with federal units on fugitive searches. ![](/images/longeye-logo-dark.png) ## Analysis for extremely large PDFs We've been working hard to improve support for PDFs with hundreds of thousands of pages (35MB+). This month, we introduced a new document detail page with a theme navigator that groups pages by the themes they discuss. Document chat is also much better now: each time you ask a question, the chat bar explains how it's processing your question. PDF citations now cite specific lines, not pages, and content search is much more reliable. ![](/images/blog-feb-geospatial.png) ## Saved chat history By popular demand, we finally built saved chat history. You can now revisit past chats and continue them—or start a fresh chat—right in the same spot. And, if you share with the org, your teammates can pick up where you left off. In addition, you can now save chat prompts. If you frequently send the same prompt to different files, save it once and reuse it next time. ![](/images/blog-oct-screenshot.png) ## Thematic Search is live in Audio & Video You can now search thematically through audio and video transcripts. For example, if you're looking for something like "gun," you can type "handgun" or "violence" and it will come up. Search across audio and video now also works across languages—so even if the transcript is in Spanish, you can search in English and still find it. ## Improved image captioning In user testing we are observing fewer false positives in image captioning and improved image relevance scoring. ## Case management - **Org admins can now manage users directly!** You can invite new users, reset their passwords, and reset their authenticators—all without contacting the Longeye team. - You can now override a file's relevance with a thumbs up or thumbs down. This automatically sorts it to the top or bottom of your list. - We now detect common operating system files and skip them by default. You can still view them on your evidence page and request processing, but we won't process files we recognize as common device files (like icons and application files) unless you ask. - The file upload progress indicator in the bottom-right corner is now draggable, in case you want to move it around your screen. - When you upload folders and ZIP files, you can now name them, and we preserve their folder structure to keep you organized as you work the case. - Face clusters pulled from your images are now sorted by relevance, so you don't see thousands of celebrity photos high up in your filter options. ## Other improvements - AI-generated case summaries now appear on the Case Details page, giving you an instant overview of your case as soon as processing completes. - You can now search for exact phrases in audio and video transcripts by wrapping your search in quotes. - PDF citations now spotlight the exact cited text on the page, not just the page number. - Unread indicators now show which audio and video tracks you haven't listened to yet, so you can track your progress through evidence. - We improved processing of Snapchat exports with better data labeling and classification. ## Under the hood We fixed over 300 smaller issues. We've improved ingest speeds, continued to work on reliability, and we're currently improving metadata extraction from Cellebrite UFDRs. --- ### How Investigations Break Down at Scale — and What Agencies Can Do About It (2026-02-15) Investigations rarely lose momentum because of a single mistake. More often, they slow down gradually. Detectives stay busy, but progress becomes harder to measure. Conversations shift from theories and leads to questions like, "Are the calls reviewed yet?" or "Did the extraction finish processing?" The work continues, but the case itself is no longer moving forward. That drift usually has one root cause: the volume of [digital evidence has grown](https://www.nicepublicsafety.com/resources/digital-evidence-management-system-dems-guide) beyond the workflow built to manage it. These breakdowns are not dramatic. They accumulate quietly until days or weeks have passed, and by then the impact is already visible across interviews, follow-ups, and charging decisions. ## Evidence Volume Outpaces Sequential Review Cases still begin in familiar ways, a burglary, an assault, a shooting. What breaks workflows is what arrives next. If you work investigations today, you've opened inboxes containing hours of body-worn camera footage, multiple device extractions holding years of messages and app data, social media warrant returns spanning thousands of pages, and unlabeled folders of images and video clips. [National Institute of Justice](https://nij.ojp.gov/topics/forensics/digital-multimedia-evidence) guidance confirms digital evidence now appears in nearly every crime type, not just cyber cases. Traditional sequential review made sense when evidence fit one folder. Today the caseload math no longer works: - One detective typically manages 15-25 active investigations - Average case requires 45 hours of digital review - Manual workflows collapse under that sustained volume When investigators spend most of their time just opening files and navigating formats, less attention reaches the connections that actually solve cases. ## Late Context + Fragmentation Creates Drift You've probably experienced this: a key detail surfaces weeks after your first interview. A confession buried deep in call lists. You pick up a cold case that's been worked by four different detectives - all retired. A timeline contradiction hiding in thousand-page PDFs. Images that change everything, discovered after witnesses are already debriefed. [Police1 reports](https://www.police1.com/what-cops-want-2023/articles/survey-results-indicate-urgent-need-for-comprehensive-workload-analysis-and-service-delivery-revamp-F2ixCHtt8jZUmtZA/) investigators now describe their work as "digital triage," routinely skipping evidence portions just to maintain case velocity. Early decisions get made on partial information: - Interviews proceed without critical background - Follow-ups target incomplete leads - Prosecutors receive technically accurate but fragile case updates None of this reflects a lack of effort. When evidence review lags behind case momentum, every downstream step rests on incomplete context. Cases don't stop, they drift. ## Fragmentation Turns Verification Into Parallel Work The real challenge isn't evidence complexity. It's fragmentation across disconnected systems. A jail call timestamp aligns with an interview statement. A screenshot contradicts a suspect claim. A video shows a location mentioned in a transcript. [Investigators know these connections matter](https://nij.ojp.gov/topics/forensics/digital-multimedia-evidence), but tracking them manually across separate tools exhausts cognitive bandwidth. Verification becomes its own investigation. Digital forensics programs report analysts spend most time on these steps: - Relocating the exact insight moment - Confirming surrounding context - Validating original source file - Documenting clearly for prosecutors Supervisors notice when updates slow. Prosecutors notice when sourcing requires extensive reconstruction. Cases lose momentum not because evidence doesn't exist, but because defending it takes longer than finding it. ## Capacity Declines While Workloads Grow Investigative teams face evidence growth from one direction and capacity pressure from every other. [PERF's 2023 staffing survey](https://www.policeforum.org/staffing2023) found sworn officers [down 5%](https://www.policeforum.org/staffing2023) from 2020-2022, resignations up nearly 50% vs pre-2019 levels. [Cellebrite's 2024 digital forensics survey](https://www.police1.com/police-products/investigation/evidence-management/cellebrites-2024-industry-trends-survey-reveals-urgent-digital-challenges-in-law-enforcement) showed 67% reporting growing workloads, 52% expecting worsening backlogs. Municipal budget pressures compound this, [expenditures now outpace revenues](https://www.nicepublicsafety.com/resources/digital-evidence-management-system-dems-guide), leaving investigative vacancies unfilled and digital forensics under-resourced. The result: more data arriving faster than the hours available to process it. ## Agencies Adjusting Through Workflow Changes [Leading agencies](https://nij.ojp.gov/topics/articles/improving-collection-digital-evidence) aren't solving this through longer hours. They're changing how work advances from intake forward. Common patterns among teams keeping pace: - Triage immediately - context builds as evidence arrives, not after full datasets complete - Relevance-first - focus on key people/times/locations vs opening files by load order - Unified formats - audio/images/docs/video form one narrative view early - Source-linked insights - notes tie directly to original files for instant verification These changes protect what investigators do best. Interviews sharpen with better context. Follow-ups happen sooner because leads surface earlier. Prosecutors receive clearer updates with stronger sourcing. The workflow finally matches the evidence scale. Leaders frame the goal simply: reduce data-heavy administrative time, maximize judgment and community knowledge where they matter most. But any new tools must meet [investigator standards](https://www.justice.gov/olp/media/1381796/dl): traceable insights, defensible conclusions, security compliance. Next, we'll examine what investigators need from AI tools to trust them in mission-critical workflows, focusing on verification, defensibility, and CJIS-aligned security that holds up under courtroom scrutiny. --- ### February Product Update (2026-02-01) We're showing your evidence some love this February. Here's what's new this month. ## [Evidence.com](http://evidence.com/) Integration with Axon You can now import files directly from Axon [Evidence.com](http://evidence.com/) into Longeye. Folder structures are preserved, and files stream directly to your case without manual downloads. This feature is available by request—just let us know if you'd like it enabled for your organization. ![](/images/blog-mar-case-reports.png) ## Find the Right Images, Faster - **Search by AI relevance notes** — When our AI analyzes your images, it notes why each one might matter—"shows suspect," "potential weapon," "document evidence." You can now search these notes directly to find exactly what you need. - **Hide small images** — One toggle filters out icons, thumbnails, and other visual noise so you can focus on meaningful evidence. ![](/images/blog-feb-phone-data.png) ## New Filters in Explorers - **Filter calls by phone number** — Multi-select filter on the Calls page lets you quickly find all recordings involving specific phone numbers. - **Filter by file type** — New dropdown on the Evidence page lets you show only PDFs, audio files, images, or other types. - **Sort audio/video by duration** — Find short clips or long interviews instantly with the new duration sort option. ![](/images/blog-jan-financial.png) ## Smarter AI Chat - **Chat scope toggle for social media** — Choose whether you want the AI to focus on just the current conversation or search across your entire case. - **Custom formatting** — Ask for summaries in your preferred style—bullet points, narrative, or anything else. - **Copy AI responses** — Easily copy any summary or chat response to paste into your reports. ## Audio, Video & Documents - **Copy transcripts** — New button lets you copy the full transcript with one click. Works with translated transcripts too. - **Word documents viewable as PDFs** — We now convert .doc/.docx files automatically, preserving comments so you can view them directly in your browser. - **PDF titles in chat** — Document chat now shows the actual PDF title instead of the filename. - **Large PDFs load smoothly** — Massive PDF reports (like Cellebrite exports) now load in your browser with page-by-page navigation, so you can scroll through thousands of pages without waiting. ## Navigation & UX - **Redesigned Home page** — Two-column layout keeps your case metrics visible while you explore themes and highlights. - **Redesigned Image filters** — We moved the filters on the Image page into their own column on the left side so that you can focus as much screen space as you'd like on viewing the images themselves, as we continue to make Image analysis even better with more filters in the future. - **Deep-linking from Files page** — Click the audio, video, or image badge on any file to jump directly to that item in the explorer. - **Incident alerts** — If we're experiencing issues, you'll see a status banner in the sidebar. - **Password-protected files** — Now clearly labeled instead of showing confusing error messages. ## Wins from the Field Every month, investigators across the country use Longeye to break cases faster. Here are a few recent highlights: - **South San Francisco PD** — During an on-site training, a detective uploaded 25 hours of jail calls from an attempted homicide case. Within 5 minutes, Longeye surfaced a Spanish-language call where the suspect instructed his girlfriend to coordinate alibis with his co-defendant's girlfriend. - **Oklahoma DOC** — An investigator uploaded PayPal and Green Dot records from a fraud victim. Longeye identified the payments that matched the exact amounts and timing the victim had described, saving hours of manual review. - **Oklahoma City PD** — A Spanish-speaking officer had reviewed jail calls and reported nothing relevant. When the detective ran the same calls through Longeye, the AI found a confession buried in just a few seconds of audio that had been missed, which provided critical evidence in a case with no DNA. **Have a win to share?** We'd love to highlight your work. [Submit your story here →](https://forms.gle/2F4bwGCnZNVb1PA56) ## Tip from Investigators **Export PDF reports from Cellebrite.** When you export a Cellebrite extraction as a PDF report, Longeye can search through text messages, emails, call history, and more—all at once. It's a fast way to surface what matters across an entire phone dump. ## Under the Hood We fixed **45+ bugs** this month, including face bounding boxes not appearing on images, citation links breaking in social media chat, and large ZIP files (100GB+) failing to extract. **Performance:** Evidence page search is now near-instant thanks to a new database index—no more waiting when you have thousands of files. **Reliability:** Large file uploads can now be resumed if they're interrupted. Just re-select the same file to pick up where you left off. **Security:** We patched 2 authentication vulnerabilities and added encryption across additional storage services to keep your data protected. --- ### January Product Update (2026-01-01) Happy New Year from the Longeye team! We've been hard at work making Longeye faster, smarter, and more reliable for you. In the last month, we've shipped over 170 improvements to the app, ranging from new features to performance enhancements to critical bug fixes. ## New Features ## Face Detection & Clustering Longeye now automatically detects and clusters faces across all images in a case. This means you can: - **Filter by face** — Click on a face thumbnail to instantly find all images containing that person - **See face clusters** — View how many photos each person appears in ("this person is in 15 photos") - **Label people** — Name your clusters for easy reference (hover and edit inline) - **Face bounding boxes** — See exactly where faces are detected in the lightbox view ## Chat with Text Files You can now chat with plain text files (.txt, .log) using AI! Ask questions about the document and get answers with clickable citations that highlight the exact text being referenced. Perfect for quickly navigating lengthy documents. ## Smarter Audio & Video Summaries Our new two-pass AI summary system generates more accurate summaries of audio and video transcripts. The AI first drafts a comprehensive summary, then reviews it against the original transcript to fact-check and improve accuracy — all with timestamped citations you can click to jump right to that moment. ## Delete Evidence Files Case admins can now permanently delete evidence files. Select multiple files at once, confirm the deletion, and they're removed immediately from your case. (Files are retained in storage for 7 days before permanent deletion, just in case.) ## Read-Only Case Access New **Case Viewer** role lets you give users read-only access to specific cases. Viewers can browse evidence and chat with files, but can't upload or modify anything. Great for supervisors or external reviewers who need visibility without edit access. ## Deep-Linking from Evidence Page to Explorers Click the new enrichment badges on the Files page to jump directly to the Audio, Video, Images, or Documents explorer with that file in focus. No more hunting through pages to find what you're looking for! ## Bug Fixes - **PDF processing** — fixed some parsing failures for special characters - **Investigator notes** can now be saved even when very long (was hitting a size limit) - **ZIP file extraction** handles edge cases better with improved compression ratio calculations - **Multi-tab sessions** no longer conflict during file uploads - **Audio & translation improvements** — better speaker identification and improved language translation accuracy. Citations on the summaries for audio and video files are now faster to load. That's a wrap on the end of 2025! We're so excited to be back at it in 2026 with fresh energy and well-rested brains. As always, find us at [support@longeye.com](mailto:support@longeye.com) if you have questions or run into any issues. --- ### The Hidden Costs of Manual Evidence Review (2025-12-15) If you spend time around investigators, a pattern quickly stands out. They rarely complain about the dangerous parts of the job or the unpredictable hours. What wears people down are the long stretches of digital review that no one outside the work ever sees. Hours spent inside outdated audio players and document viewers, watching transcripts, chat logs, and exported data load at a snail's pace. Nights spent combing through tens of thousands of pages that still leave unanswered questions. Manual evidence review used to be manageable. Today it is one of the most time-consuming and mentally draining parts of investigative work, and many agencies are just coming to recognize how deeply it affects both investigators and case timelines. ## Evidence Piles Up Faster Than Agencies Can Review It Across agencies, investigators describe review sessions that look less like structured analysis and more like digital triage. A detective may start the evening with dozens of audio clips, a multi-thousand page PDF, a device extraction filled with years of messages, and a folder of images that may or may not be relevant. All of it opens slowly. None of it is connected. Each piece requires separate tools and separate attention. This is not an edge case. It is increasingly normal in many jurisdictions. To understand the scope of the shift, consider the pace of evidence growth: We know this problem is universal. A [study from the National Institute of Justice (NIJ) found that the digital evidence required per case increased by 80% in one year alone](https://nij.ojp.gov/topics/articles/digital-evidence), demonstrating just how fast this volume is expanding. The evidence changed faster than the workflow, and investigators were left to absorb the impact without additional time, tools, or staffing. ## Manual Review Slows Investigations Before Anyone Notices Cases rarely break all at once. They slow down in small, subtle ways and begin to drift. A detective sees the files that load fastest instead of the ones that matter most. A prosecutor receives an update that is technically correct but incomplete. These small delays stack quickly and compound across the life of a case. When review cannot keep up with the pace of incoming evidence, early decisions are made on partial information. By the time critical details surface, the best window to use them may have already passed. ## Evidence Overload Drives Burnout and Attrition Detectives expect long hours and difficult subjects. What they do not expect is how isolating digital review can become. Listening to disturbing audio alone at the end of a shift. Sorting through graphic or emotionally heavy screenshots. Switching between tools that do not talk to each other. Trying to keep track of timelines while moving across formats that were never meant to align. Much of this happens after hours because investigators need to spend the day contacting suspects and supporting victims. But the mental load accumulates. The fear of missing something important is always there. It is a quiet pressure that rarely appears in official documentation but has real effects on morale. This intense cognitive burden directly impacts staffing and agency health: The impact of this burnout is clear. A [2023 survey from the Police Executive Research Forum (PERF) highlighted that the high mental and administrative workload is a driving factor in the retention crisis, with 47% of agencies reporting that the number of resignations increased in the previous year](https://www.policeforum.org/staffing2023). This is why many leaders recognize that the solution lies in smarter tools, not just harder work: As the former President of the International Association of Chiefs of Police (IACP), Chief Steven R. Casstevens, noted, ["It's imperative that we equip our officers with technology that not only enhances safety but also reduces the administrative and cognitive burden, allowing them to focus on community engagement and complex investigations"](https://www.theiacp.org/iacp-technology-center). The math bears this out as well. We recently saw a homicide case with 850+ hours of jail calls and 50,000+ pages of PDF documents to review. Even if a detective worked 2x as hard, there's no way to get through that while other cases keep coming in. We need to give investigators access to products that enable them to move 100x as fast. ## Manual Review Wastes High-Value Investigative Time The cost of manual review is not just the number of hours spent. It is the value of the hours being spent. Though salaries vary widely by region, the average detective costs an agency roughly one hundred and sixty-seven thousand dollars annually when salary, benefits, equipment, and training are included. Prosecutors and public defenders cost even more. These are some of the most precious hours in the justice system. When those expensive hours are consumed by tedious sorting, the case itself suffers: This loss of focus has measurable consequences. ["Every minute an investigator spends on menial, manual sorting is a minute they aren't spending building the case narrative or interviewing a key witness," stated a technology lead quoted in Police Chief Magazine, "and that directly impacts the likelihood of solving the case"](https://www.policechiefmagazine.org/improving-clearance-rates-in-investigative-case-management/). When investigators spend days watching, listening, and sorting through material that does not require detective-level judgment, the agency loses time it can never recover. Cases move slower. Staffing pressure increases. And the people doing the work feel the strain most directly. ## Agencies Are Shifting Workflows to Keep Up with Evidence Scale The agencies keeping pace with Big Evidence are not doing it through longer hours or larger teams. They are adjusting their workflow. They start review earlier. They begin with relevance, not file order. They bring formats together sooner, so audio, images, documents, and video strengthen each other. They expect every insight to tie back to its source immediately. These adjustments give detectives time back. Interviews improve. Case updates become clearer. Follow-up work happens faster. The job regains its rhythm. Investigators did not choose for review to take over their work. It simply grew to fill the space as evidence multiplied. Agencies that recognize this and shift their workflow are already seeing cases move faster and teams experience less strain. The ones that wait will feel the pressure grow each year. The volume is not going to shrink, but the approach to handling it can evolve. And when it does, investigators get to return to the parts of the job that truly require their judgment, experience, and presence in the community. In the next article, we'll look at a question every agency is now facing: if traditional workflows are falling behind, can new tools like AI help investigators keep up — and what would it take for them to trust those tools? — *Guest contributor* [*Antoine Sawadogo*](https://www.linkedin.com/in/antoine-sawadogo-53b8252a3/) *studies criminology and human rights at the University of Lille, with experience supporting judges, court systems, prosecutors, and defense attorneys.* --- ### December Product Update (2025-12-01) Hi! We hope you're all staying warm and cozy out there. Our team has been hard at work making Longeye faster, smarter, and more reliable for you. In the last month, we've shipped over 200 improvements to the app, ranging from new features to performance improvements to bug fixes. ## Mobile Device Forensics Made Easy **Cellebrite UFDR files are now supported.** Upload forensic extractions from mobile devices and Longeye will now automatically parse all their contents into the Explorers, just like with ZIP files. ## Better Transcription Quality You'll now see quality warnings when Longeye is unable to confidently parse a transcript—so you know when to trust the AI transcription and when to listen to the original audio. We added detection for transcript gaps, as well as a fallback audio model, to improve the quality of our transcript coverage. ## Image Metadata and Annotations Photos now extract EXIF metadata when available, and display that GPS information on an interactive map: - GPS location on an interactive map with accuracy radius - Timestamp with timezone information - Device make and model (iPhone 14, Pixel 8, etc.) Images now also support tagging & comments, so you can group like images together, leave notes for yourself and your teammates, and export grouped images for your reports extremely easily. ![image tagging & comments](/images/blog-dec-image-tagging.png) ## More New Capabilities - **Word-by-word highlighting for audio and video.** Words now highlight in real-time on the audio and video playbacks, making it much easier to follow along on transcripts. This even works while following the English translation of transcripts in other languages. - **Timestamp extraction from audio and video filenames.** We added support for a bunch of new date formats found in the names of audio and video files. If your files have timestamps in their names, but we're not yet extracting them, just let us know! - **Date filters for audios and videos.** You can now filter to the specific time of occurrence for audio and video files. - **Relevance filters on every page.** You can now filter by relevance in every single Explorer. - **Large PDF chat.** We're working hard to support chat and parsing for increasingly large PDFs. You can now chat with PDFs with thousands of pages, and we're working to support PDFs with hundreds of thousands of pages as well and more. - **Increased support for Video and Image file formats.** You can now browse HEIC image files directly in Longeye. We support playback for an increasing number of video formats and rendering for an increasing number of image formats. - **Files process sooner.** Data ingestion will now start processing your files the moment we receive the first one, as opposed to the moment we finish receiving the batch. This gives you your first results in just minutes, even when uploading very large batches of data. ## Bug Fixes that Mattered We squashed many bugs that were blocking your workflows, particularly around loading complex file formats and very very large files. We gave a lot of TLC this month to our file transcripts, and made the transcript quality, playback, and citations all much more transparent for you. ## What's Coming? Some things that we're particularly excited about in December include supporting increasingly large PDF files, image clustering around similar objects & people, and expanding the scope of chat to your entire case as opposed to individual files. Stay tuned for updates on these things and more in our next newsletter in January. We hope you all enjoy a wonderful and peaceful holiday season. If you have any questions or feedback, don't hesitate to reach our team at [support@longeye.com](mailto:support@longeye.com). --- ### Is Your Agency Keeping Up With Big Evidence? (2025-11-15) If you work investigations today, you have probably felt the shift even if no one has named it. A warrant return arrives as a multi gigabyte PDF instead of a few pages. A batch of jail calls appears in your inbox before you have finished the last one. A phone extraction stalls your computer. A single case generates more digital material in a week than older cases generated in months. These moments add up to one clear reality. Evidence load did not just increase. It exploded past the limits of the workflow built to handle it. If you feel those things, welcome to the era of Big Evidence. ## What Big Evidence Actually Means Big Evidence is not just a large collection of files. It is the point where the volume, variety, and speed of digital evidence surpass what a traditional human centered workflow can reasonably process. It is evidence that grows faster than your available hours. It is evidence that arrives fragmented across many formats. It is evidence that contains the answers but hides them inside overwhelming scale. Digital evidence is also now present in about [90 percent of criminal cases](https://www.policechiefmagazine.org/building-stronger-bonds-osint-digital-forensics-magnet/), according to recent reporting in Police Chief Magazine. The average case has five devices and [60 hours of material to review](https://www.police1.com/police-products/investigation/investigative-software/why-ai-and-cloud-solutions-are-no-longer-optional-for-police-investigations). And as a result, [70% of investigators tell Police1](https://www.police1.com/police-products/investigation/investigative-software/why-ai-and-cloud-solutions-are-no-longer-optional-for-police-investigations) they don't have enough time to go through all their digital evidence. That means nearly every investigation today involves digital material that is more complex and more demanding than workflows originally designed for it. And investigators aren't keeping up. Most agencies are already operating in this reality, even if they have not started calling it by name. ## What Big Evidence Looks Like in Real Cases Across the country, the size of digital evidence has expanded far beyond what most investigative teams were built to manage. Industry research highlights [steady growth in body worn cameras, in car video systems, and digital evidence management platforms](https://ijis.org/wp-content/uploads/2024/05/LEITTF-Digital-Evidence-Whitepaper.pdf). Agencies are generating more video, audio, and digital media than many back end systems were originally designed to support. On the ground, that trend shows up in cases like these examples from real cases worked with Longeye: - More than three thousand hours of jail calls spread across four thousand files in a cold case review - Sixteen thousand pages in a single social media warrant return - A one terabyte device extraction filled with years of messages, photos, and app data - Hundreds of multilingual calls needing rapid triage In some departments, the daily output of calls alone can consume more than a year of listening time for a single investigator. This pattern appears across agencies of every size, from small municipal departments to large county sheriff offices. The trend is unmistakable. Evidence volume is rising faster than the capacity to handle it. ## How Cases Lose Momentum Cases rarely stall all at once. They drift. A detective opens the files that load fastest instead of the ones that matter most. An interview moves forward because the witness is available before the core context arrives. A prosecutor receives an update that is accurate but incomplete. These delays stack quietly. After a few days, the case is no longer moving on the timeline it needs. The best windows for follow up work narrow. Leads go cold. Big Evidence changes timing, and timing shapes outcomes. ## The Strain Investigators Carry The shift also changes the day to day experience of investigative work. More evidence means more screens, more time spent inside large documents, more audio reviews late in the day, and more device exports that need several attempts to load. Less time supporting victims and contacting suspects. Investigators push through all of it because they care about their cases, but the unstructured volume adds pressure. It takes time away from the core work they are trained to do. Verification becomes harder. Cognitive load increases. None of this shows up in reports, but it affects the job every day. Nobody wants to hear echoes of jail calls in their sleep. ## Why This Problem Is Not Going Away Every trend but one in digital evidence points in one direction. Devices store more. Cloud platforms archive deeper histories. Messaging apps generate denser threads. Body cameras record longer sessions at higher quality. Courts continue requesting more documentation, not less. And as Police1 recently wrote, [the explosion of affordable video technology is rapidly increasing the number and types of digital evidence sources, overwhelming many back end systems](https://www.police1.com/vision/ai-ready-policing-enhancing-law-enforcement-computer-systems-for-tomorrows-challenges). The one trend that isn't pointing toward growth? Investigator capacity. Some agencies have seen their jurisdictions double in population since the last time they added detective headcount. The workflow has to evolve. ## How Agencies Are Beginning To Adapt Many agencies are already making small but meaningful adjustments that help them keep pace with Big Evidence. They begin processing evidence immediately so the case starts moving as soon as files arrive. They prioritize relevance instead of file order. They bring formats together early so that audio, images, documents, and video reinforce the same narrative. They link findings to their sources right away so verification stays close to discovery. These changes give investigators more time for the tasks only humans can do. Interviews improve. Case updates become clearer. Follow up work happens sooner. The job feels more manageable again. Investigators everywhere are adapting to this new environment, often without calling it a shift. They are finding ways to keep their cases moving in a world where the evidence no longer fits the old workflow. Agencies that acknowledge this reality early are already seeing the benefits. Others will feel the strain grow each year. Big Evidence is here to stay, but the approach to handling it can evolve. And when it does, investigators regain the time and focus they need to work cases the way they were meant to. At Longeye, we specialize in helping agencies combat their Big Evidence challenges. If you are interested in a free consultation to assess the extent to which Big Evidence impacts your operations, we can help you with a quantitative baseline exercise to share with your command staff. Feel free to request your no-strings-attached consultation here. — *Guest contributor* [*Antoine Sawadogo*](https://www.linkedin.com/in/antoine-sawadogo-53b8252a3/) *studies criminology and human rights at the University of Lille, with experience supporting judges, court systems, prosecutors, and defense attorneys.* --- ### Longeye + Microsoft Entra ID: SSO, CJIS, and Federated Authentication (2025-11-01) ## Introduction: Strengthening CJIS Compliant Authentication Large parts of the Criminal Justice Information Security (CJIS) policy revolve around authentication and verifying that a user is who they claim to be. Historically, this meant "one set of credentials per system," which required different username and password combinations for every system an officer needed to access. Detectives may be superhuman in many dimensions, but their appetite for remembering dozens of passwords is normal. With many personal and workplace credentials already in use, adding yet another agency specific password is not something most officers look forward to. ## Why Single Sign On (SSO) Matters for Agencies Single sign on (SSO) addresses this challenge by allowing officers to log in with one username and password, along with MFA, and then use that authentication seamlessly across multiple systems. A modern police department typically wants: - One secure, centrally managed login per officer - MFA delivered through a CJIS compliant hardware authenticator - The ability for IT staff to provision or deprovision access in a single place This is especially valuable when officers retire, transfer, or change roles. Managing access across many systems manually is time consuming and prone to errors, which is something CJIS standards aim to avoid. ## The Rise of Federated Authentication Consumers are familiar with "Login with Google" or "Login with Facebook." Agencies can accomplish the same approach, but with their own private identity provider (IDP) or broader identity and access management (IAM) system. Historically, Microsoft Active Directory has been the most widely used IAM. In Microsoft Azure, this system is now known as Microsoft Entra. Entra provides the tools an agency needs to manage secure access to computing resources. ## Longeye's Approach: Seamless Integration with Existing IAM Systems Longeye's authentication system is built on the Okta cloud based IAM platform. While Okta can manage its own users and credentials, it also supports federation, which is the key enabling technology behind SSO. When properly configured, federation is straightforward to set up. Longeye's security team partners with an agency's IT staff to integrate their SAML (Security Assertion Markup Language) or OIDC (OpenID Connect) IDP, and the process is usually completed in less than an hour. ## Benefits of Federated Authentication for Agencies Federated authentication offers several advantages beyond convenience. ### 1. Simplified Identity Proofing Identity proofing, which verifies a user's identity using government IDs, does not need to be repeated. If a user is already proofed in the agency's IAM, they are automatically validated for Longeye. ### 2. No Coordination Needed for User Provisioning New users can be assigned to Longeye inside the agency's IAM. There is no need to contact Longeye or manage another account. Two key reasons make this simple: - Longeye has no per user licensing charges - Users can log in at their convenience once access is assigned ## Microsoft Entra ID Is Only the Beginning Entra is the first federated IAM system Longeye officially supports, but agencies across the country use a variety of IAM platforms. Longeye is committed to integrating with whatever systems agencies rely on to deliver fast and secure access. SSO reduces onboarding friction and provides detectives with a simple and fully CJIS compliant authentication experience. It is not often that increased security results in a smoother user workflow, but this is one of those cases. --- ### October Product Update (2025-10-01) Hi, we hope you're enjoying the spooky season! The Longeye team was in Denver last week for IACP, and it was great to see so many of you in person, make new connections, and hear your excitement and enthusiasm for how Longeye is helping with your investigations. *Here's what's new in Longeye this month:* ## Chat with Documents Have you ever had to go through a long PDF document with pages of emails, or maybe a CSV with extensive text messages? Now, these files will show up in the Documents explorer with a "Chat" icon next to them. Interrogate these files like you do your audio, video, and images — with deep-linked citations of course. ![](/images/blog-jan-chat.png) ## Document Relevance Scores & Improved PDF Parsing Just like in the other Explorers, Longeye now automatically scores each document based on its content's relevance to the case details you provided. Longeye can flag connections and contradictions across your evidence, and we have improved our PDF processing to also read scanned documents and handwritten notes in your case details. ## Faster Processing Processing is now dramatically faster. After your files finish uploading, you'll start seeing results within minutes instead of waiting for entire batches to complete. ## Support for All Video Formats When we first rolled out support for video analysis last month, it only worked for .mp4. Now, we support *all* video formats — .mov, .avi, .mkv, .3gp, .webm, .wmv, .flv, .m4v, .mpeg, .mgp, etc! We're looking forward to hearing about what you find. ## Smarter Audio Analysis We've made major improvements to transcription accuracy and readability: - *Case-specific vocabulary:* Longeye now extracts case-relevant keywords for better accuracy with proper nouns and slang - *"Sounds like" matching:* The system is now trained to allow for less-precise matching, so that you can be certain you don't miss a reference, even when audio quality is poor - *Improved readability:* Long conversations are now broken up into smaller 20-second segments with intelligent splitting at sentence boundaries, making transcripts easier to follow along when one person is talking for a long time ## Improved Image Explorer We added a pan & zoom capability to our image explorer and improved its layout so you can easily see each picture's relevance score, relevance explanation, caption, and file name — right alongside the image itself. ## Activity Log For your IT and compliance teams, we've added an audit log page directly in the app. This will tell you about all core events, like when a new user is added to your account, when new evidence has been uploaded, or when someone's role has changed. ## Upload Improvements We've made file uploads more flexible and reliable: - *Drag and drop from anywhere:* Upload files from any page within a case - *Stay logged in during large uploads:* No need to wiggle your mouse while uploading large bodycam files - *Better validation:* The Case Details page now redirects to evidence when files are uploaded to it that are only supported for evidence analysis ## Under the Hood Under the hood, we've been working hard to make initial file processing faster and more reliable, and we have improved your ability to see how file processing is going on the Evidence page with a "Processing files..." indicator. We've also released dozens of small improvements based on your feedback at IACP, including better keyboard shortcuts, clearer role selection, and visual polish across the application. --- Questions or feedback? We'd love to hear from you! You can reach us at team@longeye.com. The Longeye Team --- ### September Product Update (2025-09-01) Hi there! We've been really busy making Longeye even better for your investigations. Here's what's new this month: ## Smarter AI Analysis We're constantly working to improve how our AI understands your case context. This month, we added a new Case Details page to collect Case Reports, Incident Reports, and other documents with context on the details you may be looking for (see below). Our AI now looks for these details, like physical descriptions of people, cars, and weapons, in every new piece of evidence you upload. The system will highlight these for you by ranking them with higher relevancy scores. ## New Case Details Page When working a new case in Longeye, you want to get our AI engine up to speed as quickly as possible. Previously, you'd have to type out the details of what you were looking for in our case description area. But now, you can simply drag & drop your Case Report, Incident Reports, and other contextual information into the Case Details page. We'll will extract the necessary context from those documents and start looking for what you're looking for *automatically*! ## Video Evidence Support Ever need to search bodycam footage for the moment someone said something? Ever have an interview room video you needed transcribed and summarized? Now you can upload and analyze MP4 video files! Automatically transcribe videos and translate their audio into English. Watch moments in videos as you listen to their audio tracks, summarize the salient points in seconds, and inquire about who said what & when via Chat. Now you can rip through the key moments of videos just like you do with audio files. ## New Documents Page As we talk to detectives across the country, we keep discovering new types of documents they have to work with to close cases. Today, we are taking the first step towards being able to handle any type of document by organizing these into our new Documents page. We're looking for your feedback on how you'd like to interact with documents! We will be adding text extraction, translation, and "Chat with file" to PDF documents in the coming weeks. ## Granular Access Control Is your IT team all over you for for making sure the system stays CJIS compliant? Granular user roles are now available at both the Case and Organization level. You can set users to be Admins, Members, and Viewers. Longeye's Role Based Access Control (RBAC) features make sure you are always in full control of who can interact with your data, from 911 calls to the courtroom floor. Find these new controls on the Settings pages for your Cases and Organization. ## ZIP File Processing Warrant returns often come in the form of ZIP files, sometimes even ZIPs of ZIPs, and it can be hard to find the files worth your time. Well... there's no need to unzip these anymore! Just drag them into Longeye as ZIPs and we'll take care of it. We support up to 1000 files per ZIP archive, and files up to 10GB in size. We are always looking for ways to streamline your work! ## Case Bookmarks Your caseload is massive, and you can't realistically make progress on all your cases at once. We introduced bookmarks on your case list, so you can make sure your work in Longeye reflects the urgency of your work in the community. Thanks for being part of the Longeye community! We're continuously working to make digital evidence analysis faster and more intuitive for investigators like you. Questions or feedback? We'd love to hear from you! The Longeye Team --- ## Legal ### Privacy Notice Longeye's Privacy Notice describes how we collect, use, and protect personal data. Topics covered include: our role in processing personal data, collection and use of personal data, additional uses, disclosure of personal data, privacy choices, children's personal data, security, and contact information. Last updated August 26, 2025. Full text available at: https://www.longeye.com/LongeyePrivacyNotice.pdf ### Cookie Notice Longeye's website uses only strictly necessary cookies for hosting infrastructure. We do not use analytics, advertising, or third-party tracking cookies. Last updated March 27, 2026. Full text available at: https://www.longeye.com/cookie-policy