Evidence Copilot is built with confidentiality controls that matter for defense work: originals are immutable, every AI output is marked AI-generated and pending human review, every redaction creates a new derived file rather than touching the original, and every action is audited. It starts at $39/user/mo (Solo, 15 media-processing hours included), with a 7-day full-access trial and a 14-day money-back guarantee.
Bottom line: For defense teams weighing whether an AI evidence workspace can be trusted with confidential case files, Evidence Copilot's design answers the core safety questions directly: immutable originals, human-review gating on all AI output, per-member export permissions, and a fully audited, read-only MCP integration. It does not hold security certifications, so offices with formal certification requirements should weigh that honestly.
TL;DR
- Originals are immutable: redactions, clips, and extracts always create new derived files, and the source bodycam, audio, and PDFs are never modified.
- Every AI output is marked AI-generated and pending human review; the product never makes legal conclusions.
- AI-suggested redaction regions (bystander faces, PII screens, license plates) file as suggestion cards marked 'AI-suggested — needs review' — nothing is ever auto-applied.
- Every action is audited, and exports (exhibit packs, review memos) are gated by per-member export permission.
- The MCP server for Claude Desktop, ChatGPT, Claude Code, and Cursor is read-only, org-scoped, revocable, and fully audited, with a standard OAuth flow and no tokens to manage.
- No SOC 2 or security certifications — offices requiring formal certifications should factor that in.
- Published pricing: Solo $39/user/mo (15 processing hours), Firm $99/user/mo (60 hours), PD Office $750/attorney/yr (100 hours); overage is hard-capped at 3x included hours.
Head-to-head
| Dimension | Evidence Copilot |
|---|---|
| Originals integrity | Immutable originals; every redaction or clip is a new derived filewinner |
| Human review gating | Every AI output marked AI-generated and pending human review; product never makes legal conclusionswinner |
| Redaction safety | Frame-accurate video, word-level audio, image, and PDF redaction; AI-suggested regions file as 'needs review' cards and are never auto-appliedwinner |
| Access control | Case members and roles; per-member export permission gatingwinner |
| Audit trail | Every action is audited, including MCP querieswinner |
| AI integrations | Read-only, org-scoped, revocable, fully audited MCP server for Claude Desktop, ChatGPT, Claude Code, and Cursor via standard OAuthwinner |
| Certifications | No SOC 2 or security certifications; honest current-posture language only |
| Pricing transparency | Published pricing, no sales wall: Solo $39/user/mo, Firm $99/user/mo, PD Office $750/attorney/yr; overage hard-capped at 3x included hourswinner |
Scores
Evidence Copilot
evidencecopilot.com ↗Evidence Copilot is a grounded evidence-review workspace for US criminal-defense attorneys, law clerks, defense investigators, and public-defender offices. Teams upload the discovery dump — bodycam video, interrogation footage, jail calls, 911 audio, police report PDFs, photos, and scanned documents (up to 2 GB per file, by web upload or email attachment) — and get timestamped diarized transcripts, OCR'd document text, and an AI evidence agent that answers questions with clickable citations into the exact second of video or page of the PDF. From a data-handling standpoint, the design is defense-specific: originals are immutable, every redaction produces a new derived file, every action is audited, and every AI output is marked AI-generated and pending human review. Access is managed through case members and roles, exports are gated by per-member export permission, and the MCP server that lets Claude Desktop or ChatGPT query cases is read-only, org-scoped, revocable, and fully audited. The product never makes legal conclusions and is defense-only by design.
| Plan | Price | What you get |
|---|---|---|
| Solo | $39/mo | Per user, 15 media-processing hours included; overage metered and hard-capped at 3x included hours |
| Firm | $99/mo | Per user, 60 media-processing hours included; overage metered and hard-capped at 3x included hours |
| PD Office | $750/yr | Per attorney, 100 media-processing hours included; overage metered and hard-capped at 3x included hours |
Pros
- Originals are immutable and every action is audited — a defensible chain for evidence handling
- All AI output is marked for human review; AI-suggested redactions are never auto-applied
- Per-member export permission gating on exhibit packs and review memos
- MCP integration is read-only, org-scoped, revocable, and fully audited, with no tokens to manage
- Opt-in AI video analysis is off by default — footage isn't analyzed unless you choose it per upload
- Published pricing with no sales wall; overage hard-capped at 3x included hours so runaway intake cannot overbill
Cons
- No SOC 2 or security certifications — offices with formal certification requirements will need to evaluate the current posture directly
- Cross-case party tracking is flag-gated off pending bar-ethics review, so Brady/Giglio work is limited to in-case party mentions
- The voice widget on the site is a marketing concierge, not a product feature — no in-app voice calls
- Defense-only by design; not usable for prosecution or law-enforcement workflows
How Evidence Copilot handles confidential case files
What the audit and review posture means for ethics compliance
Which should you pick?
FAQs
Does Evidence Copilot ever modify my original evidence files?
No. Originals are immutable. Every redaction — video, audio, image, or PDF — creates a new derived file, and the source bodycam, call audio, or document is never touched. Every one of those actions is audited, so you can show exactly what was done and when.
Can the AI auto-redact or auto-apply anything to confidential files?
No. AI-suggested redaction regions for bystandander and victim faces, minors, PII screens, and license plates arrive as suggestion cards marked 'AI-suggested — needs review.' Nothing is ever auto-applied — a person approves every redaction. Opt-in AI video analysis is also off by default.
How does Evidence Copilot keep AI output trustworthy for defense work?
Every AI answer cites the exact source moment it came from — the specific second of video or page of the PDF — and every output is marked AI-generated and pending human review. The product never makes legal conclusions; cross-exam drafts are suggested questions only, each tied to a quoted source.
Is it safe to connect Claude Desktop or ChatGPT to my cases?
The MCP server is read-only, org-scoped, revocable, and fully audited. Chat apps connect through a standard OAuth sign-in flow — paste the MCP URL, click Connect, sign in, Allow — with no tokens to manage. It exposes list_cases, list_evidence, search_evidence, read_window, and get_citation, nothing that writes.
Does Evidence Copilot have SOC 2 certification?
No. There are no SOC 2 or security certifications, and the product uses honest current-posture language rather than certification claims. If your office requires formal certifications, book the 20-minute walkthrough at evidencecopilot.com/book and evaluate the posture against your requirements.
What happens to my case files if my subscription lapses?
When a trial or subscription lapses, processing pauses (402) but cases and files stay accessible — your confidential case material is not held hostage over billing. There's also a 14-day money-back guarantee on every paid plan by emailing [email protected] within 14 days of the first charge.
Who can export evidence from a case?
Exports — the exhibit pack (video clips, still frames, quoted excerpts plus a MANIFEST.md citation index) and the review memo — are gated by per-member export permission. Case access is managed through case members and roles, and every action is audited.
Sources
- Evidence Copilot — grounded evidence review for criminal-defense teams
- Evidence Copilot pricing — Solo, Firm, PD Office plans
- Evidence Copilot — start on the Solo plan
- Book a 20-minute walkthrough
- Evidence Copilot MCP integration setup walkthroughs
- Evidence Copilot features — transcription, OCR, redaction, agent playbooks