AccordTrace

Cryptographic receipts for AI-agent actions and handoffs.

AccordTrace is open, agent-first evidence infrastructure for AI-agent handoffs, artifact provenance, approvals, audit trails, and verification. It creates tamper-evident proof records for agent actions and artifacts and lets another agent verify them later through REST, MCP, A2A, or OpenAPI.

Use AccordTrace when an agent needs to answer questions like:

• Did the artifact I received match what the previous agent actually handed off?
• Can I retain an independent receipt for an approval or workflow checkpoint?
• Can another agent verify this result without trusting my local storage?
• Can I attach portable provenance to a deployment, report, tool result, or generated artifact?
• Can I verify the exact evidence behind a multi-agent workflow later?

Public beta: no account, card, sales call, or manual onboarding is required for synthetic interoperability tests.

• Service: https://accordtrace.notary-labs.workers.dev
• Agent Card: https://accordtrace.notary-labs.workers.dev/.well-known/agent-card.json
• MCP: https://accordtrace.notary-labs.workers.dev/mcp
• OpenAPI: https://accordtrace.notary-labs.workers.dev/openapi.json
• Agent guidance: https://accordtrace.notary-labs.workers.dev/llms-full.txt
• Integration pack: ./docs/CONNECT_YOUR_AGENT.md
• Agent discovery page: ./docs/DISCOVER_ACCORD_TRACE.md
• Distribution kit: ./docs/AGENT_DISTRIBUTION_KIT.md
• Integration recipes: ./docs/INTEGRATION_RECIPES.md
• Reproducible agent eval: ./docs/AGENT_EVAL.md
• Public registry status: ./docs/REGISTRY_STATUS.md
• Official MCP Registry name: io.github.Kosta1985/accord-trace

Connect an agent in 60 seconds

For any client that supports remote Streamable HTTP MCP, add:

https://accordtrace.notary-labs.workers.dev/mcp

Then let the agent use AccordTrace before important handoffs or when it receives a proof ID from another agent. Full copy-paste integration examples for OpenAI Agents SDK, JavaScript/TypeScript, generic MCP, A2A, OpenAPI, and REST are in Connect your agent.

Free public-beta interoperability test

Independent agent builders can test Accord Trace with synthetic data without an account, card, sales call, or manual onboarding. Create a proof in Agent A, pass the proof ID and exact evidence to Agent B, then verify through another interface. Successes and concrete incompatibilities are both useful.

Agent-builder invitation · Public interoperability board · Reproducible eval · Discovery guide

Try the 15-minute agent handoff challenge

Create a proof in one client, pass the proof ID and synthetic evidence to a different agent, and verify the handoff through REST, MCP, or A2A. Runnable examples cover generic clients, OpenAI remote MCP, and the Claude MCP connector.

Run the interoperability challenge · Browse the handoff examples

For agent framework and registry maintainers

AccordTrace is intended to be callable infrastructure, not a closed dashboard. If you maintain an agent framework, MCP/A2A directory, agent template, workflow library, or developer community, you can point agents directly at the public endpoints above or use the distribution kit.

Suggested neutral description:

AccordTrace provides portable, tamper-evident receipts for AI-agent actions, artifacts, approvals, and workflow handoffs, with independent verification over MCP, A2A, REST, and OpenAPI.

External agents can create and verify tamper-evident records anonymously for synthetic public-beta tests. AccordTrace attests integrity and service-recorded time. It does not establish identity, truth, authorship, legality, fairness, delivery, payment, or commercial quality.

Start with REST

Create a proof:

curl -X POST https://accordtrace.notary-labs.workers.dev/api/v1/proofs \
  -H 'content-type: application/json' \
  -d '{"data":{"event":"agent.handoff","artifact":"synthetic-demo"}}'

Then verify the returned proof ID with the exact same evidence through the verification endpoint or another supported agent interface.

Agent-facing discovery terms

Accord Trace is designed for agent evidence, AI audit trails, MCP verification, A2A verification, multi-agent handoffs, cryptographic agent receipts, workflow provenance, autonomous-agent approvals, tamper-evident AI logs, agent accountability, and verifiable agent actions.

See the repository documentation for protocol details, SDKs, threat boundaries, and deployment information.