Evidence workspace
Versioned workspaces, exportable audit trails, local-first by default. Compare runs without surrendering control of your records.
Why this exists: EU AI Act Article 50 transparency duties have applied since 2 August 2026; the relevant Annex III high-risk transition is 2 December 2027 (Annex I: 2 August 2028). NIST AI RMF 1.0 is voluntary and IMDA's 22 January 2026 agentic framework is non-binding guidance. These dates and product mappings need qualified legal review. Every tool below is free to run. Paid report delivery and checkout are paused pending legal, pricing, and fulfillment review.
Where do I start? — 3 questions
Answer yes or no. We point you at the diagnostics that match your actual exposure, so you don't run all thirteen on day one. Nothing is stored — it runs in this tab.
01Do any of your agents act — spend money, send messages, deploy, delete, or write to a system?
02Do any of your agents make decisions about people — hiring, credit, access, pricing?
03Do you ship AI-assisted content into the EU or to regulated clients?
Your governance file — browser-local
Coverage dashboard
Every diagnostic you finish on this site is recorded in an unkeyed, hash-chained ledger in this browser — nothing is uploaded. This panel shows which of the 13 diagnostics you have actually run, what they scored, and exports a governance file you can file or hand to a reviewer.
Paid concepts · not available for purchase
Six ways to turn evidence into action
These are research proposals, not current features or offers. Checkout remains paused; no data leaves your browser through these concepts unless a future implementation explicitly says so.
Continuous drift watch
Bring-your-own CI or scheduled checks flag changes against your own baselines; configurable thresholds, signed exports, and a one-click pause.
Incident rehearsal kit
Facilitated scenarios, role cards, timed decisions, and debrief actions make response practice repeatable—not a compliance verdict.
Evidence integrations
Open CSV, JSON, and webhook adapters connect issue trackers and GRC tools. Inspect every field and revoke access at any time.
Approval & exception register
Explicit owners, expiry dates, rationale, and immutable export history make exceptions visible and easy to review or withdraw.
Reusable stakeholder packs
Transform selected evidence into editable board, engineering, or broker packs with source links, uncertainty labels, and accessible charts.
Design bar: high utility, leverage, agency, integrity, trust, coherence, quality, performance, resilience, composability, accessibility, adaptability, and clarity. No dark patterns, surprise telemetry, or automated legal conclusions.
Seasonal Tools & Multi-App Showcase
12 bundled tools + 1 sold separately, grouped by use
Free tools are available now. Paid checkout is paused; hosted subscriptions are not available. Catalogue prices below are proposals, not live offers. Paid files are general self-guided workbooks, not personalized reports.
Filter by category. Every card shows the same price and opens the same tool page as the full catalogue. Categories are a way to browse, not a change to what each tool costs or includes (12 bundled tools below + BIAS/14 sold separately).
Showing 12 of 12 apps.
PROVENANCE/CHECK$9/mo
Content Disclosure
Real client-side EXIF/C2PA image provenance inspector + EU AI Act disclosure-readiness diagnostic.
Open Tool →
ATTEST/LOG$29/mo
Authorization
Browser-local, hash-chained authorization notes. Not a digital signature or independent proof of approval. Hosted Pro is not available.
Open Tool →
PERIMETER/14$79
Supply Chain
14-field exposure score for agent-facilitated supply-chain incidents. Free scorecard, paid remediation report.
Open Tool →
ATTRIBUTE/CHECK$49
Incident Attribution
Scored 14-field accountability readout mapped to Arts. 12, 14, 26, 50, 72–73 — self-assessment, no model inference.
Open Tool →
READY/14$297
Org Readiness
Scores agent-readiness across ownership, workflow, observability, and governance. Paid tier is a self-guided remediation workbook with a general 90-day planning framework.
Open Tool →
POSTMORTEM/14$29
Content Engine
Templates branded breakdowns of documented public AI failures — or your own incident notes — for LinkedIn, newsletter, or client advisory memo.
Open Tool →
GATE/CHECK$19/mo
Code Review
Light/dark PR verdict tool — scores whether a PR was reviewed with real judgment or rubber-stamped, 7-point rubric.
Open Tool →
LEASH/14$89
Insurability
Combined agent containment score framed against ISO CG 40 47/48 generative-AI CGL exclusion endorsements. Free score, paid renewal-disclosure draft for broker review.
Open Tool →
METER/PROOF$29/mo
Billing / Metering
Browser-local, unkeyed hash-chained usage ledger. Free cap: 15 events. Not a signature; hosted Pro is not available.
Open Tool →
SETTLE/14$89
Dispute readiness
Evidence, authority, insurance-notice and remediation readiness before an AI-agent dispute escalates.
Open Tool →
ANNEX/14$99
Classification
EU AI Act classification worksheet covering system category and provider/deployer roles.
Open Tool →
DRIFT/14$79
Production drift
Model, prompt, tool, policy, data and behavior-drift control assessment.
Open Tool →
New — sold separately, not in the bundle
BIAS/14$129
Bias & Fairness
14-field bias-exposure diagnostic for agents that touch people — training-data provenance, subgroup evaluation, decision explanation, review path, and the written evidence you'd need if a decision is challenged.
Open Tool →
THE STACK/HUB BUNDLE
All 12 bundled tools shown above (plus 1 sold separately: BIAS/14) — proposed workbook bundle; the treatment of subscription tools is undecided. Not available to purchase.
Separately: $810 one-time + $86/month
$349
GET THE STACK →