Trooth Network MCP¶
A company's witnessed trust record, answerable from any MCP client. Trooth runs a remote, read-only MCP server that lets an AI assistant check a company's Trust Profile, take a neutral read of a domain's public security surface, verify a signed Trust Ledger Token, or ask about Trooth itself. Every answer names its source and date; witnessed evidence is labeled apart from what a company declares, and unsettled subjects come back marked unknown rather than guessed.
Server type: Hosted, remote (Streamable HTTP), public and read-only
Endpoint: https://api.trooth.co/public/mcp
Auth: none (no account, no token)
Registry: io.github.trooth-eng/trooth-network
Tools: trooth_public_trust_profile, trooth_outside_in_read, trooth_verify, trooth_ask
Built by: Trooth (trooth.co)
Why This Matters for Operators¶
Vendor and partner checks are normally a private-investigator workflow. Trooth turns the public trust layer into live agent context.
First, provenance is structural. Witnessed Trust Profiles and Trust Ledger Tokens are signed evidence; outside-in reads are live neutral observations. The server returns honest absences for unknown subjects instead of inventing a rating, which matters when an agent's answer feeds a procurement decision.
Second, zero friction. The endpoint needs no account or token, so any MCP client (ChatGPT custom connector, Claude connector, Cursor, Cline, VS Code) can add it in one paste.
Third, it is a different data class. Identity, security, privacy and AI-practice signals with source labels complement the financial and operational connectors an agent already has, giving diligence answers a second axis.
Tools and Capabilities¶
| Tool | What it returns |
|---|---|
trooth_public_trust_profile |
A company's witnessed Trust Profile or Network standing, with provenance |
trooth_outside_in_read |
A live, neutral read of a domain's public security surface |
trooth_verify |
Re-verify a signed Trooth Trust Ledger Token |
trooth_ask |
Ask about Trooth (products, methodology, pricing) |
Verification (Aug 31, 2026)¶
- Live JSON-RPC initialize verified: POST to
https://api.trooth.co/public/mcpreturnedtrooth-mcp v1.1.0with tools/resources/prompts capabilities and instructions describing the read-only trust layer. - README documents the endpoint, the four tools, and official MCP Registry record
io.github.trooth-eng/trooth-network.
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.