A2Cn

I built an open protocol for Agent-to-agent commercial negotiation

Last verified:

Visit A2Cn

What is A2Cn?

A2CN is an open, neutral protocol for agent-to-agent commercial negotiation across organizational boundaries. It addresses the missing layer between agents that can talk and pay, but cannot yet safely commit to binding commercial terms. The site positions A2CN as the negotiation layer for B2B procurement, renewals, and supplier negotiations, where two agents from different companies need a shared way to exchange offers, verify authority, and create tamper-evident records.

The protocol’s core value is standardization for commercial deal-making between autonomous agents. A2CN defines eight components: discovery, mandate verification, session invitation, offer exchange, deal-type terms, session state machine, transaction record, and audit log. Those pieces are meant to cover the full lifecycle of an automated negotiation, from finding the other agent to producing a signed record of the final agreement.

A2CN also includes implementation and integration support. The website says the project ships a Python reference implementation, platform adapters, an MCP server, and an LLM negotiation skills file. It is presented as something procurement platforms, contract AI tools, and agent frameworks can adopt without major platform changes.

The site is aimed at teams building procurement systems, contract automation products, and agent frameworks that will need cross-company commitment workflows. It also targets design partners who want to pilot the protocol and help shape later versions. In short, A2CN is for organizations trying to make agent-driven commercial negotiation safe, verifiable, and interoperable.

A2Cn pricing

Pricing model: Freemium

The website does not show a paid pricing page or subscription tiers. It presents A2CN as Apache 2.0 licensed, with the spec and reference implementation available for use, and it invites design partners to pilot the protocol. The site also mentions a planned Meeting Place hosting layer for future version 0.3, but says it is not built yet; no free tier versus paid plan breakdown is listed on the site.

A2Cn pros

  • Open protocol for agent negotiation
  • Neutral standard across organizations
  • Focuses on binding commercial terms
  • Covers procurement and renewals
  • Includes mandate verification
  • Uses cryptographic offer signing
  • Supports strict turn-taking
  • Defines clear session state machine
  • Produces tamper-evident transaction records
  • Includes structured audit logs
  • Has deal-type schemas
  • Supports extensible custom terms
  • Provides reference implementation
  • Offers platform adapters
  • Includes MCP server support
  • Compatible with common agent frameworks
  • Has bilateral negotiation walkthrough
  • Apache 2.0 licensed
  • Designed for cold-start adoption
  • Built for enterprise procurement use cases

A2Cn cons

  • Still v0.2.0
  • Meeting Place not built yet
  • TypeScript reference implementation planned
  • SDKs not yet shipped
  • A2A extension proposal still in review
  • Requires cryptographic and DID complexity
  • Focused on B2B negotiation only
  • Limited to current supported deal types
  • Needs design-partner adoption to mature
  • Not a full procurement platform
  • No centralized registry by design
  • Integration depends on adapter coverage
  • Primarily for organizations already using agents
  • Author is a solo builder
  • Some features are roadmap items

Frequently asked questions about A2Cn

What is A2CN?

A2CN is an open protocol for agent-to-agent commercial negotiation. It is designed for situations where AI agents from different organizations need to negotiate and commit to business terms such as procurement deals or SaaS renewals.

What problem does A2CN solve?

It solves the gap between agents that can communicate and agents that can actually agree on binding commercial terms. The website says there is currently no neutral protocol for verifying authority, exchanging offers, and keeping a tamper-evident agreement record across company boundaries.

What are the main protocol components?

A2CN defines eight components: discovery, mandate verification, session invitation, offer exchange, deal-type terms, session state machine, transaction record, and audit log. Together, they handle discovery, authorization, negotiation flow, final record creation, and compliance output.

How does A2CN verify authority?

The protocol uses mandate verification with cryptographic proof that an agent has authority to commit its organization. The site mentions W3C DIDs and a two-tier model using Declared and Verifiable Credential methods.

What kinds of deals is A2CN built for?

The website highlights goods_procurement and saas_renewal as normative deal types. It also says the protocol supports extensible custom_terms, so other commercial negotiation formats can be modeled as well.

Does A2CN keep a shared record of the deal?

Yes. A2CN creates a dual-signed transaction record that is generated independently by both parties. The site emphasizes that neither side controls the authoritative copy, which makes the record tamper-evident and shared.

What integrations are available?

The site says A2CN includes adapters for Fairmarkit, Keelvar, and Salesforce Revenue Cloud, plus an MCP server that works with Claude Desktop, LangChain, LangGraph, Agentforce, and Cursor. It is positioned as something that can be dropped into a procurement platform with zero platform changes.

Is there an implementation I can use now?

Yes. The website says there is a Python reference implementation with 202 passing tests. It also says the protocol spec is available and the code is Apache 2.0 licensed.

Is A2CN fully finished?

No. The website describes A2CN v0.2.0 as shipped, while the A2A extension proposal is still in review. It also lists a future Meeting Place hosting layer, a TypeScript implementation, and SDKs as planned work.

Who is A2CN for?

A2CN is aimed at procurement platforms, contract AI companies, agent framework builders, and design partners that need cross-organizational commercial commitment. The site also mentions people working on LLM-based agents who want a negotiation skills file or protocol layer for safe deal-making.

Categories

Use cases

Browse all AI tools on NeedAnAI