Trooth OS
Install once. Trooth keeps your page true.
Trooth OS evaluates your systems against 10 frameworks, signs every claim, anchors the audit chain daily, and publishes your Public Trust Profile. Install once. We run it from there.
- 10 frameworks
- Signed claims
- Daily anchor
- Public Trust Profile
How it works
Built on a global edge runtime.
We evaluate your systems against 10 regulatory frameworks on a continuous schedule. Every claim is signed with a two-signature notary token. The audit chain anchors to a public time-stamp authority every 24 hours. You see it all in a browser.
Capabilities
Inside Trooth OS.
Evidence Engine
Deterministic, weighted classifiers across 10 frameworks. Every result traces to the observation behind it.
Two-Signature Notary
An independently auditable signed token. Your inner signature plus our witness signature.
Tamper-Evident Audit Chain
Forward-secure. Replay any slice on demand.
Daily Time-Stamp Anchoring
A single daily anchor to a public time-stamp authority.
Witnessed Record
Every category re-checked on each scan, each result traced to the observation behind it. Your Public Trust Profile updates automatically.
Public Trust Profile
Every customer gets an SEO-indexable profile. Free.
Included capability
EU AI Act compliance, built into Trooth OS.
EU AI Act automation is part of the platform, not a separate product. Trooth OS classifies your systems against the Act's risk classes, generates the disclosures the regulation requires, and keeps watching as the rules evolve. The rules move. We track them. You don't have to.
Trooth automates. Trooth never signs on your behalf.
EU AI Act API referenceRisk classification
Automatic mapping of every system to the EU AI Act risk classes, prohibited, high-risk, limited, and minimal, recomputed on every scan.
Article 50 disclosures
Generate the transparency disclosures the Act requires for AI systems that interact with people, ready to publish.
Continuous regulatory monitoring
As the Act's guidance and harmonized standards change, Trooth OS tracks the deltas and flags what newly applies to you.
The signed chain, stage by stage.
You grant Trooth permission to look. Write scopes are not requested, so Trooth can never change the systems it reports on.
Illustrative sample. The real run happens against your systems, and every line it writes is signed.
Architecture
Every hop, and what it is unable to do.
Topology
↺ The whole path repeats on a schedule, which is why a profile does not go stale.
Your stack
Cloud accounts, identity providers, repositories and endpoints, connected with read-only grants you approve one at a time.
Cannot
Cannot be written to. No write scope is requested, so Trooth is unable to change the systems it reports on.
Payload shape
{ "connector": "aws", "grant": "read-only", "scopes": ["describe*", "list*", "get*"], "write_scopes": []}Shapes are real; the values are an illustrative sample, not anyone's account.
Proof
What the audit chain records.
- 07:12:43CONTINUOUS_SCANACTIVE
- 07:12:48EVIDENCE_CAPTUREPASS
- 07:12:55SIGNATURE_CHAINSECURE
- 07:13:02AUDIT_ANCHORANCHORED
- 07:13:11FRAMEWORK_COVERAGEWITNESSED
- 07:13:25PUBLIC_PROFILEINDEXED
- 07:13:35WITNESSED_RECORDUPDATED
Illustrative sample. Not live data.
For developers
Drop it into your stack.
CLI, GitHub Action, embeddable badge, or a single API call. Free and open-source, no account required to try the scanner.
Drop it into your stack.
Five ways to read one witnessed record: a command, a pull-request check, a line of HTML, a single GET, or a connector your assistant reads. The scanner is report-only and needs no account, so you can run it before you decide anything.
Trooth Network is a listed app in the ChatGPT directory. In Claude, add it as a read-only connector. No key, no account.
# Add Trooth as a read-only connector (MCP)# ChatGPT: open Trooth Network in the app directory# Claude: Settings > Connectors > Add customconnect https://api.trooth.co/public/mcp# Report-only preflight from the terminaltrooth scan plan.json# .github/workflows/trooth.yml- name: Trooth preflight (advisory) uses: troothllc/trooth-action@v1 with: api-key: ${{ secrets.TROOTH_API_KEY }} plan-file: plan.json fail-on: none # report-only<div id="trooth-trust-badge" data-slug="your-co"></div><script src="https://trooth.co/badge.js"></script># Returns the signed, public Trust Profile (use your own slug)curl https://api.trooth.co/public/trust/your-coPricing
Get the full platform. Included free.
Works with the rest of Trooth.
One account, one witnessed evidence chain. Every product below reads from it.