Longeye
AI investigative workspace for justice agencies to analyze digital evidence faster and more efficiently.
Website: https://www.longeye.com
Cover Block
From the public record
| Field | Value |
|---|---|
| Name | Longeye |
| Tagline | AI investigative workspace for justice agencies to analyze digital evidence faster and more efficiently. [Longeye, October 2026] |
| Headquarters | San Francisco, US [Longeye, October 2026] |
| Founded | 2025 [Longeye, October 2026] |
| Stage | Series A [Business Wire, October 2026] |
| Business model | SaaS |
| Industry | Legaltech |
| Technology | AI / Machine Learning |
| Geography | North America |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (2) |
| Funding label | $10M+ |
| Total disclosed funding | ~$25,000,000 [Inc., October 2026] |
Links
From the public record
- Website: https://www.longeye.com/about
- Careers: https://www.longeye.com/careers
The Short Version
PUBLIC Longeye builds an AI investigative workspace for justice agencies, and it has become investable to watch because it paired a 2025 founding with more than 50 reported agency users and a $20 million Series A by October 2026 [Business Wire, October 2026] [Longeye, October 2026]. The company’s pitch is straightforward: help investigators review large volumes of digital evidence across documents, phone records, financial transactions, audio, and video, while keeping findings tied back to source material, a control point that matters in criminal-justice workflows [Longeye, October 2026] [Forbes Councils].
The founding story is early but coherent in public materials. Longeye was founded in 2025 in San Francisco, and its leadership combines public-safety interest on the CEO side with senior platform-engineering experience on the CTO side [Longeye, October 2026] [Forbes, October 2019] [Datanyze]. Guillaume Delpine is identified as co-founder and CEO, with Longeye describing him as a former Skydio leader, while Danielle Man is identified as co-founder and CTO, with prior senior engineering leadership at Apollo GraphQL [Longeye, October 2026] [LinkSV] [Datanyze].
The product differentiation, based on available evidence, rests less on a generic chatbot interface than on evidence ingestion and analysis across mixed data types, plus traceability and controlled model deployment for sensitive legal workflows [Rippling] [Forbes Councils] [Inc., October 2026]. Inc. reported that investigators can query the system in plain English to extract evidence, build timelines, and compare witness accounts, while company materials say the underlying AI runs in a private closed system on Amazon Bedrock and does not use public-defender case materials to train models [Inc., October 2026] [Longeye, October 2026]. Those claims are directionally attractive, but most of the technical-control detail remains company-sourced.
On funding, the public record is unusually clear for such a young company. Longeye says Andreessen Horowitzs American Dynamism Fund led a $5 million seed round in 2025, and Business Wire reported that N47 led a $20 million Series A in October 2026, bringing disclosed funding to roughly $25 million, consistent with Inc.s reporting [Longeye, October 2026] [Business Wire, October 2026] [Inc., October 2026]. The business model is SaaS, and current hiring for enterprise and mid-market account executives selling into police and sheriffs departments suggests the company is building a direct go-to-market motion around agency budgets rather than a services-heavy model [Rippling].
What to watch over the next 12 to 18 months is less the headline funding and more the shape of adoption. Public evidence points to breadth, with more than 50 justice agencies using the platform and customer agencies reportedly having used it across more than 2,300 investigations, but the market still lacks named customer references, disclosed contract sizes, retention data, and independent validation of the companys speed claims [Business Wire, October 2026] [Citybiz, October 2026] [Rippling]. If Longeye can convert early agency count into repeatable deployments, referenceable outcomes, and trusted procurement footholds across both law-enforcement and defense-side users, the category fit will look more durable than the current company narrative alone establishes.
Single-source, plausible -- Mix of independent reporting from Business Wire, Inc., and Citybiz, with several material product and team details still sourced primarily from company materials and job postings.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Series A |
| Business Model | SaaS |
| Industry / Vertical | Legaltech |
| Technology Type | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | $10M+ (total disclosed ~$25,000,000) |
The Company in Brief
PUBLIC
Longeye is early, but the basic company record is unusually crisp for a 2025 startup. The company describes itself as a San Francisco-based business founded in 2025, focused on an AI investigative workspace for justice agencies [Longeye, October 2026]. Public materials identify Guillaume Delépine and Danielle Man as co-founders, and a LinkedIn profile identifies Quentin Delepine as a co-founder as well [Longeye, October 2026] [LinkedIn].
The milestone sequence is straightforward in the sources. Longeye says Andreessen Horowitz's American Dynamism Fund led a $5 million seed round in 2025, with participation from Seven Stars Capital [Longeye, October 2026]. By October 2026, the company announced a $20 million Series A led by N47, bringing disclosed funding to $25 million according to Inc. [Business Wire, October 2026] [Inc., October 2026]. That pace suggests a company moving quickly from formation to institutional backing, although the public record here is still weighted toward company and announcement materials rather than independent operating disclosures [Longeye, October 2026] [Business Wire, October 2026].
Single-source, plausible -- Confirmed primarily by the company website, with partial corroboration from Business Wire and Inc.
What They Have Built
Mixed sourcing
What stands out first is not a broad AI claim but the shape of the workflow Longeye says it is addressing: large evidence sets that arrive in different formats and need to remain usable in a legal process. The company describes the product as an AI investigative workspace for justice agencies, built to help users analyze documents, phone records, financial transactions, audio, and video while preserving links back to source evidence [Longeye, October 2026]. In press coverage, the product is also described as letting investigators query evidence in plain English through a chatbot interface to extract evidence, build timelines, and identify discrepancies across witness accounts [Inc., October 2026].
The public positioning suggests the wedge is evidence review and organization rather than a generic assistant layer. A sales posting says the platform can review heterogeneous evidence types including text, images, audio, video, financial records, and telephone records, and claims analysis can run "up to 100x faster" than manual review, although that speed figure appears to come from company-controlled materials rather than independent testing [Rippling]. Separate company-linked materials say findings remain traceable to original sources and that the AI is designed for criminal-justice requirements, both of which matter if the system is being used in investigations that may later face courtroom scrutiny [Forbes Councils].
Longeye also makes a specific privacy and deployment case, again mainly through its own materials. The company says its AI runs in a private, fully closed system on Amazon Bedrock and says public-defender case materials are never used to train large language models, either its own or those of providers [Longeye Public Defenders page, Retrieved 2026]. That architecture detail should still be treated cautiously until corroborated through a verified technical demo or third-party security documentation, but it is directionally consistent with the product's intended buyers and the trust requirements of criminal-justice workflows. A Member of Technical Staff role in San Francisco points to active in-house engineering buildout, though any deeper stack inference would remain speculative from the public record [Rippling].
Unconfirmed -- This section relies materially on company-controlled sources and job postings, with limited independent corroboration from Inc.
Market Size and Demand
From the public record
Digital evidence volume, rather than a lack of investigators, appears to be the immediate market pressure point here, and that is the clearest reason this category matters now [Inc., October 2026] [Citybiz, October 2026]. Public reporting on Longeye does not cite a third-party TAM, SAM, or SOM study for AI software sold into justice agencies, so the market has to be framed more conservatively through observed demand signals and adjacent public-sector software budgets rather than a headline market-size figure [Business Wire, October 2026] [Inc., October 2026]. What is visible is narrower but still meaningful: as of October 21, 2026, Longeye said more than 50 justice agencies were using its platform, and Citybiz reported those agencies had already worked on more than 2,300 investigations spanning documents, images, audio, video, and structured records [Business Wire, October 2026] [Citybiz, October 2026].
The demand case rests on workload complexity across multiple evidence types. Longeye's reported usage metrics point to review volumes of more than 15 million pages of documents, 10 million images, 32,000 hours of audio and video, and more than 1 billion rows of financial and telephone records, which suggests the relevant market is not generic legal software so much as investigative analytics for high-volume digital discovery inside criminal-justice workflows [Citybiz, October 2026]. Inc. separately described the product as a tool to help police sift through evidence, including by asking questions in plain English, extracting evidence, constructing timelines, and identifying discrepancies between witness accounts [Inc., October 2026]. That set of tasks places Longeye at the intersection of e-discovery, digital forensics workflow, records analysis, and case-preparation software, with substitution pressure likely coming from manual review, point tools, and incumbent evidence-management systems rather than from one cleanly bounded category [Inc., October 2026] [Longeye, October 2026].
A simple way to size the opportunity from the outside is to separate the directly evidenced served market from the broader analogous markets that surround it.
| Market frame | What the evidence supports | Source |
|---|---|---|
| Directly evidenced current footprint | More than 50 justice agencies using the platform as of October 21, 2026 | [Business Wire, October 2026] |
| Evidenced workload through customers | More than 2,300 investigations processed across customer agencies | [Citybiz, October 2026] |
| Analogous market: e-discovery and digital investigation workflow | Adjacent, but no third-party market-size figure was captured in the source set for this report | [Inc., October 2026] [Citybiz, October 2026] |
| Analogous market: public-safety and justice-agency software | Relevant buyer environment, but no cited budget or category-size benchmark was captured in the source set | [Business Wire, October 2026] [Longeye, October 2026] |
from the available evidence is that market validation is stronger on pain intensity than on published category sizing. Investors can see a real workload and a real buyer group, but the source set does not yet support a defensible top-down market model.
Several tailwinds are visible even without a formal market report. The first is data proliferation: the reported mix of documents, media, phone records, and financial records implies that justice agencies are dealing with multimodal evidence stacks that are hard to review with linear, human-only workflows [Citybiz, October 2026]. The second is usability: Inc.'s description of a plain-English investigative interface matters because public-sector teams often adopt software more readily when it compresses training time and reduces dependence on specialist analysts [Inc., October 2026]. The third is buyer breadth. Longeye's own materials and related coverage point not only to law enforcement, but also to prosecutors and public defenders, which broadens the adjacent budget pools if the product can satisfy chain-of-custody, privacy, and evidentiary standards across those constituencies [Citybiz, October 2026] [Longeye, October 2026].
Regulatory and macro forces cut both ways. On one side, criminal-justice agencies face pressure to process evidence faster without compromising traceability, and Longeye's materials emphasize links back to source evidence, secure self-hosted AI, and a closed system on Amazon Bedrock, all of which map to procurement concerns around auditability and data handling [Forbes Councils] [Longeye Public Defenders page, Retrieved 2026]. On the other side, the same environment is likely to slow sales cycles and raise scrutiny. Inc.'s framing, asking whether the AI can be trusted, is a reminder that adoption in this market turns not just on productivity claims but on evidentiary reliability, disclosure obligations, and whether agencies are comfortable introducing AI into investigative and case-building workflows [Inc., October 2026]. That does not negate the opportunity, but it does suggest the addressable market will expand only to the extent vendors can satisfy legal-process requirements, not simply because the underlying technical problem is large [Inc., October 2026] [Forbes Councils].
Single-source, plausible -- This section relies on named public sources for observed adoption and workflow volumes, but no independent third-party market-sizing report was captured in the source set, so TAM framing is necessarily indirect.
Who Else Is Fighting for This
MIXED Longeye is positioning itself less as a general AI assistant and more as a workflow layer for justice agencies that need to search, organize, and trace large bodies of evidence across documents, images, audio, video, financial records, and phone data [Longeye, October 2026] [Rippling].
The practical comparison set splits into three buckets, even though the company has not publicly named direct rivals. First are incumbent justice-tech and digital-forensics systems already embedded in law-enforcement workflows, which typically own record systems, evidence management, or chain-of-custody processes rather than AI-first investigative review. Second are newer AI vendors selling summarization, search, and document analysis into regulated public-sector environments. Third are adjacent substitutes, namely manual analyst review, outside investigative support, and general-purpose foundation-model tools that can answer questions over evidence sets but do not inherently provide justice-specific controls or evidentiary traceability [Inc., October 2026] [Forbes Councils]. In that context, Longeye's positioning appears narrow by design: it is trying to win the investigative workspace layer rather than the entire public-safety software stack [Longeye, October 2026].
The edge visible in public materials is product focus plus early distribution momentum, not yet a demonstrated platform moat. On product, the company consistently describes support for heterogeneous evidence types and emphasizes that findings remain linked back to source material, which matters more here than raw model capability because justice users need auditability and evidentiary confidence, not only speed [Longeye, October 2026] [Forbes Councils]. On distribution, more than 50 justice agencies using the platform by October 2026 is meaningful for a company founded in 2025, particularly in a market where procurement is slow and trust compounds gradually [Business Wire, October 2026]. That said, most of the claimed differentiation, including secure self-hosted AI, Amazon Bedrock deployment, and 100x speed claims, is still company-sourced or lightly corroborated, so the durable advantage looks perishable unless usage turns into referenceable customers, renewal proof, and integrations that make displacement harder [Longeye Public Defenders page, Retrieved 2026] [Rippling] [Inc., October 2026].
The main exposure is that Longeye does not appear to control the system of record or the broad procurement channel. Incumbents with existing agency contracts could add AI review features into evidence-management or records platforms and compete from a stronger installed base, while horizontal AI vendors can often move faster on model iteration and price if buyers decide generic document intelligence is sufficient. Longeye also seems concentrated in a narrow buyer set, with sales roles explicitly aimed at police and sheriff's departments, which suggests execution risk if demand broadens first through prosecutors, defenders, or state-level procurement vehicles rather than local law-enforcement budgets [Rippling] [Citybiz, October 2026]. The company's reported presence in public-defender offices could broaden that footprint, but that claim remains only loosely supported in the available record [LinkedIn].
Over the next 18 months, the most plausible competitive scenario is a race between workflow specificity and platform bundling. Longeye is the likely winner if justice agencies continue to favor purpose-built investigative tools with closed-system handling, source traceability, and modality coverage across text, media, and structured records, because those requirements reward specialization over generic copilots [Forbes Councils] [Longeye Public Defenders page, Retrieved 2026]. The loser in that case would be manual review itself, which is the substitute Longeye is directly attacking with its speed and case-throughput pitch [Rippling] [Pendium.ai]. The opposite outcome is also easy to see: if incumbent public-safety vendors or broader AI vendors make justice-grade audit and security features good enough, then a standalone investigative workspace could lose ground because it sits above systems that others already own. In that version of the market, general-purpose AI tooling is the winner if agencies decide acceptable compliance matters more than purpose-built workflow depth, and Longeye is the one under pressure because its public moat is still emerging rather than settled [Inc., October 2026] [Business Wire, October 2026].
Opportunity
PUBLIC The prize here is not a better case-management feature set, but the chance to become a default investigation layer for justice agencies that now sift through fast-growing digital evidence with workflows that still look largely manual or fragmented [Inc., October 2026] [Longeye, October 2026].
Longeye's headline opportunity is to become the system of record for AI-assisted evidence review across policing, prosecution, and defense. That outcome is still early, but it is not purely theoretical on the public record. As of October 21, 2026, the company said more than 50 justice agencies were already using the platform, and Citybiz reported those agencies had used it across more than 2,300 investigations involving documents, images, audio, video, and structured records at substantial scale [Business Wire, October 2026] [Citybiz, October 2026]. The product also appears designed around a specific operational pain point rather than a broad assistant pitch: investigators can query evidence in plain English, build timelines, and connect findings back to source material, while the company emphasizes traceability and a closed system architecture for criminal-justice use cases [Inc., October 2026] [Forbes Councils] [Longeye, October 2026]. If those claims hold under deeper diligence, the reachable upside is that Longeye becomes infrastructure embedded in investigative work rather than a discretionary analytics tool.
The upside path likely depends less on a single blockbuster customer and more on a repeatable wedge that broadens across adjacent agency types.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Cross-agency standardization | Longeye starts in police and sheriff's departments, then becomes a shared investigative layer across prosecutors and public defenders working on the same evidence universe | Proof that one workflow can support multiple justice stakeholders without breaking chain-of-evidence expectations [Inc., October 2026] [Longeye, October 2026] | Public materials already place Longeye across justice agencies, not only law enforcement, and a company-linked claim says public defenders are already in the user base [Citybiz, October 2026] [LinkedIn] |
| Digital-evidence operating system | The company becomes the default front end for reviewing heterogeneous evidence sets, from documents and images to phone and financial records | Continued growth in case volumes and evidence complexity, plus product wins in search, timeline construction, and discrepancy detection [Citybiz, October 2026] [Inc., October 2026] | Reported usage spans more than 15 million pages, 10 million images, 32,000 hours of audio and video, and more than 1 billion rows of financial and telephone records, which suggests the product is already being pointed at multi-modal data at meaningful scale [Citybiz, October 2026] |
| Trust-led category leader | Longeye wins because it pairs AI productivity with controls that justice agencies can defend internally and in court | Buyer preference for traceability, closed-system deployment claims, and explicit limits on model training from sensitive case materials [Forbes Councils] [Longeye Public Defenders page, Retrieved 2026] | The company is already marketing around source traceability and private model operation rather than only speed, which fits a category where explainability and evidence provenance matter as much as automation [Forbes Councils] [Inc., October 2026] |
The compounding dynamic, if it emerges, would come from workflow depth and institutional trust rather than a consumer-style network effect. A first win inside an agency can pull in more investigators, more case types, and eventually neighboring functions that touch the same evidence base. The more the platform proves it can handle mixed evidence formats while preserving links back to original materials, the harder it may be to dislodge from day-to-day investigative practice [Longeye, October 2026] [Forbes Councils]. The early signs are directional rather than conclusive: more than 50 agencies in roughly a year after launch, plus reported use across thousands of investigations, suggest that deployment can travel beyond pilot mode if procurement and legal review are cleared [Business Wire, October 2026] [Citybiz, October 2026].
The size of the win is difficult to anchor cleanly from the public record because no verified market-size dataset or close public comparable is included in the supplied sources. Still, one scenario frame is available from financing and adoption alone. If Longeye were to convert early agency traction into a durable justice-tech platform, the company could grow far beyond the $25 million total funding reported since launch in late 2025 [Inc., October 2026]. A plausible upside case is a scaled public-safety software asset with strategic value tied to investigative workflow ownership and high switching costs, though any valuation translation here would be scenario, not a forecast, and requires evidence that current deployments expand, renew, and survive adversarial scrutiny in real casework [Business Wire, October 2026] [Inc., October 2026].
Inferred, not confirmed -- This section relies on a mix of independent coverage and company-linked claims. Adoption and funding are corroborated by named publishers, but product-control claims and parts of the cross-agency expansion thesis remain company-sourced or partially corroborated.
Sources
From the public record
[Longeye, October 2026] Longeye About | https://www.longeye.com/about
[Business Wire, October 2026] Longeye Raises $20M Series A as More Than 50 Justice Agencies Adopt Its AI Investigative Platform | https://www.businesswire.com/news/home/20261021359350/en/Longeye-Raises-%2420M-Series-A-as-More-Than-50-Justice-Agencies-Adopt-Its-AI-Investigative-Platform
[Inc., October 2026] This Startup Raised $20 Million to Help Police Sift Through Evidence. Can Its AI Be Trusted? | https://www.inc.com/aaron-mok/this-startup-raised-20-million-to-help-police-sift-through-evidence-can-its-ai-be-trusted/91419405
[Forbes Councils] Forbes Councils profile | https://councils.forbes.com/profile/Guillaume-Delepine-Co-founder-CEO-Longeye/261180e0-1a94-4267-8f02-7f26752b16f9
[Forbes, October 2019] Leaving Business School Isn’t Easy, But Your Passion May Lay Outside Of It | https://www.forbes.com/sites/frederickdaso/2019/10/27/leaving-business-school-isnt-easy-but-your-passion-may-lay-outside-of-it/
[Datanyze] Danielle Man's email & phone | Longeye's Chief Technology Officer & Co-Founder contact info | https://www.datanyze.com/people/Danielle-Man/2331783081
[Rippling] Longeye Mid-Market Account Executive posting | https://ats.rippling.com/longeye/jobs/af12bcf3-2df9-496a-b2d7-40c3700a0fae
[LinkSV] Link Silicon Valley | https://linksv.com/company_detail?ByUrl=1&codeCompany=111332&frmMail=2
[LinkedIn] Quentin Delepine - Skydio | LinkedIn | https://www.linkedin.com/in/qdelepine/
[Citybiz, October 2026] Longeye Raises $20 Million Series A to Expand AI Investigative Platform | https://www.citybiz.co/article/922569/longeye-raises-20-million-series-a-to-expand-ai-investigative-platform/
[Longeye Public Defenders page, Retrieved 2026] Longeye Public Defenders page | https://www.longeye.com/public-defenders
[Pendium.ai] Longeye AI investigative workspace profile | https://pendium.ai
[LinkedIn, Retrieved 2026] Guillaume Long - France | Professional Profile | LinkedIn | https://www.linkedin.com/in/guillaume-long-b9137177/
Articles about Longeye
- Longeye’s AI Investigative Workspace Lands With 50 Justice Agencies — The Series A-funded startup is betting on secure, self-hosted AI to speed up case reviews for law enforcement and public defenders.