DevLandThe footage decides. Nothing can search it.
Doctrine Engine is the white-label vision and OCR engine that can. Licensed under your brand, live in 60–90 days — your customer never sees us.
One or two hours of film. Thirty-plus hours of labor.
A majority of agencies still do it by hand.
The statutes give 45 days. The backlogs grow 33%.
The hardest pages decide the matter. They come back empty.
So we built the engine that reads them.
Every finding cited to page, line, and second.
It wears your logo. It never shows its name.
The category gets decided in the next 12–24 months.
Three clocks are running at once — on your customers, on their requesters, and on your competitors.
Statutes mandate release
California's deadline for releasing critical-incident footage. New York allows 48 hours — and agencies cannot bill requesters for the redaction labor.
Requests compound
Federal FOIA requests in FY2024, up 25% year over year, with backlogs up 33%. Every one of them lands on someone's review queue.
Platforms buy, not build
K1's backing behind Reveal alone. Relativity, Exterro, OpenText and Motorola all acquired their engines rather than build them.
Cal. AB 748 · DOJ OIP 2025 · PrivSource 2026
We sell revenue, not cost.
The market already pays retail for this work, in two currencies: gigabytes and footage-hours. Wholesale sits under half of that retail. The spread is your margin, under your brand, at your price.
Documents, at retail
The market rate to process one gigabyte for review — with the hardest pages, the ones that decide matters, coming back empty.
Redaction, at retail
Manual redaction labor per footage-hour. A majority of agencies still do it by hand: thirty-plus labor-hours to clear one or two hours of film.
Wholesale, from us
Per gigabyte for documents and per footage-hour for video, tiered by volume against an annual minimum. Rate card on request under NDA.
Motorola / CDG 2022 · ComplexDiscovery 2024 · Veritone 2026
One engine, two pipelines.
You integrate once. Both lanes emit the same findings schema, and every finding carries provenance — page and line for documents, frame and second for footage.
Reads what the commodity APIs return blank.
- Degraded-scan and handwriting OCR
- Layout and table extraction, structure preserved
- Bates validation and redaction detection
- Every output cited to page and line
Clears footage without a room full of contractors.
- Face and plate detection with automated redaction
- Audio redaction on the same timeline
- Transcription and speaker attribution
- Every finding cited to the second
Sealed. Input and output are the contract. One API, one findings schema, under your brand — we never appear in front of your customer.
Deployed where the evidence may travel. An appliance on your customer's hardware — zero egress — or a sealed single-tenant instance, metered hourly. Source media is purged to zero once findings are returned.
Built to survive the hearing on its own output.
Most vision systems are built to be impressive. This one was built by litigators — first in real matters, then productized. The difference shows up in three places, and each is a question your customer's opposing counsel will ask.
Pinned model revisions and hashed weights. The same input produces byte-identical output, so a result can be reproduced months later — by you, or by the other side.
An immutable audit log from ingest to finding: what ran, which version, on which bytes, when. The record is an artifact of the run, not a report written afterwards.
Sensitive stages are gated on an explicit, logged, per-matter authorization. Absent it, the stage refuses to run — the default is silence, never a quiet inference.
Custody is an engineering property, not a badge.
Nothing here is a certification logo. Each row is a property of the pipeline, and each can be verified on your own data during the pilot.
Every file is fingerprinted with SHA-256 as it enters the pipeline. The hash anchors the audit log from first byte to finding.
On your customer's own hardware, or a sealed single-tenant instance. The appliance has zero egress; the cloud instance emits findings only.
Face and plate detection with automated redaction of frames and audio, on the same timeline as the transcript.
Source media is purged to zero once findings are returned. Pilot data is deleted within fourteen days.
Twelve to eighteen months to build. Sixty to ninety days to ship.
Hiring computer-vision engineers takes three to six months if the requisition fills, and legal-grade edge cases take longer than that. Two senior salaries are at risk before a customer ever touches the feature. The market has already chosen: Relativity, Reveal, Exterro, OpenText and Motorola all bought their engines.
| Dimension | Commodity APIs | Doctrine Engine |
|---|---|---|
| Target input | Clean, born-digital | Degraded scans, handwriting, night footage |
| Bates & redaction detection | Absent | Native, with an audit trail |
| Video evidence | Generic object labels | Redaction plus searchable metadata |
| Reproducibility | Models change under you | Pinned versions, byte-identical reruns |
| Deployment | Their cloud only | Your building, or a sealed single tenant |
| Whose brand | Theirs | Yours |
Built to wear your logo.
If your team builds an e-discovery or digital-evidence platform, this was made for you. White-label by design: your reviewers never leave your interface, your customers never hear our name, and the capability is yours to price, package and sell. You bring the platform — we disappear into it.
Your hardest data. Seven days. Zero cost.
We do not ask you to believe a benchmark. You set the metrics before we start, and your team grades the output side by side against Textract, Google and Azure on your own material.
Intake
Five hundred pages, or ten hours of footage — the material your current stack handles worst. Metrics agreed in writing first.
Scored results
Output returned against the metrics you defined, graded by your people, next to the commodity APIs on the same input.
Decision
License it, or walk away. Pilot data is deleted within fourteen days. One pilot per partner — there is never a second.
Run the engine on your footage.
Upload a bodycam or interview file. The cloud lane returns transcript segments and speaker turns, cited to the second. Cloud processing is metered; the Mac appliance runs on your own hardware at no charge.
Do not upload confidential, privileged, or case-related footage — demo material only.
Managed cloud API launches soon — today the demo runs on our appliance, free.
The Mac appliance runs on your own Apple Silicon hardware. Start the local server, then upload an MP4. No payment required.
Give us your hardest documents.
Scored results in seven days, against metrics you define, graded by your team. No cost and no obligation — if we miss your numbers, you have lost one email.