Back to directory

A2CN

An open protocol and Python reference implementation for agent-to-agent commercial negotiation, letting a buyer's and a seller's autonomous agents negotiate contract terms across organizational boundaries with cryptographic signatures instead of a shared platform.

A2CN

A2CN defines a protocol layer for commercial negotiation between autonomous agents belonging to different organizations, positioned between existing agent-to-agent communication (A2A) and payment execution (AP2) protocols, neither of which covers negotiating the terms of a deal. The reference implementation is two independent Python processes exchanging signed messages, so cryptography rather than a shared platform proves what was agreed.

Version 0.2 adds Session Invitation, a push-based handshake that lets a buyer with an A2CN endpoint invite a supplier who has not yet deployed one, using an invitation signed with the inviter's DID key over ES256 or EdDSA. Platform adapters translate third-party procurement webhooks - the repository includes one for Fairmarkit's BID_CREATED events - into A2CN's own procurement terms and back again, without requiring changes on the partner platform.

It is aimed at developers building agent-driven B2B commerce flows rather than consumer-facing negotiation tools.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.