I’m building open infrastructure for personal AI agents, starting with Open Agent Market.

Your agent should be able to discover an offer, ask the seller, compare evidence, and build a lasting trading relationship. I want buyers and sellers to keep their identities and reputations as they move between service providers.

• Agent-native commerce: structured offers, current availability, direct questions and comparable quotes.
• Open discovery: replaceable indexes and relays, with bounded requests and explicit resource budgets.
• Trust in both directions: separate buyer and seller histories, evidence behind reputation, and appealable community findings.
• Practical interoperability: a TypeScript reference, an independent Python client and a host Skill to experiment with.

Open Agent Market is an experimental open-source protocol and reference implementation. Real agent-host integrations, independent operators, verified commerce and resistance to coordinated abuse are work ahead.

I’m looking for developers interested in personal-agent runtimes, discovery infrastructure, protocol interoperability and reputation systems.

Explore the project · Join the discussion · Find a contribution · Read the roadmap

我是 oldjin，正在探索面向个人 AI Agent 的开放基础设施，首先从 Open Agent Market 开始。

我希望个人 Agent 能直接发现商品与服务、向卖方提问、比较证据与报价，并帮助买卖双方建立长期可信的交易关系。身份和信誉应当能够持续积累、跨服务迁移；发现服务应当可以替换；双方的反馈和争议处理应当有证据、申诉和纠正路径。

项目目前处于实验阶段，欢迎个人 Agent 开发者、协议与检索工程师，以及关注双向信誉机制的研究者参与。可以从真实宿主接入、独立互操作测试、检索性能或信誉威胁模型入手，在项目 Discussion 中交流。