Trust Layer
MCP server that keeps a signed, tamper-evident log of AI agent identity and actions
- Pricing
- Open source
- Platforms
- CLI, API, self-hosted
- Version
- 0.1.0
- Listing updated
- today
Similar MCP servers
About Trust Layer
Trust Layer is an open-source MCP server for security. It is available on CLI, API and self-hosted.
The README describes Trust Layer as a notary for AI agents: it records an agent's identity and each relevant action in an append-only log that can be checked cryptographically. It is built as a hash chain with signatures, so later changes to recorded entries are detectable.
The server exposes its use cases as MCP tools. The first version offers identity verification, recording auditable actions, and checking an agent's history, backed by in-memory adapters. A Postgres schema is included for a future persistent store, but the README lists integration tests against Postgres as still pending.
The project explicitly does not certify or judge whether recorded actions comply with any regulation. It suits developers who want a verifiable audit trail for agent activity and who will handle compliance decisions themselves.
Tools this server provides (3)
Taken from the project documentation; the running server may differ.
- verificar_identidad_agente
- Verifies the identity of an AI agent.
- registrar_accion_auditable
- Records an auditable action in the log.
- verificar_historial
- Verifies the recorded history.
What it does
- Tamper-evident action log
- Each recorded action is chained and signed so that alterations to past entries can be detected.
- Agent identity verification
- Verifies the identity of an AI agent before its actions are recorded.
- Audit history check
- Provides a tool to verify the recorded history of an agent.
- Domain-separated architecture
- Business logic sits in a dependency-free domain layer with ports implemented by adapters.
- Contract tests for MCP schema
- A contract test suite checks the tool schema that MCP clients see.
Good for
- Keeping a verifiable record of what an AI agent did during a workflow
- Verifying an agent's identity before it is allowed to act in a system
- Reviewing an agent's recorded action history after the fact
- Prototyping audit trails for agent deployments before adding persistent storage
Pricing
Open source
Open source, self-hosted; no paid plans mentioned.
Trust Layer FAQ
Does Trust Layer certify regulatory compliance?
No. The README states that it records identity and actions without certifying or judging whether those actions comply with any regulation.
What storage does the first version use?
The first version uses in-memory adapters. A Postgres schema is included for when a real Postgres connection is added.
How do I run the server?
The README gives the install command pip install -e ".[dev]" and starts the server with python -m src.mcp_server.servidor.
Which tools are available?
The README lists verificar_identidad_agente, registrar_accion_auditable and verificar_historial for the first phase.
Is Trust Layer free?
Yes. Trust Layer is open source, so you can use and self-host it for free.
What are the best alternatives to Trust Layer?
The closest alternatives to Trust Layer are AgentProof, Agent Audit MCP, Tracekit MCP gateway, MCP Server and MolTrust MCP Server.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.