{"data":{"slug":"agi-inc-agent-protocol","name":"agent-protocol","tagline":"Common interface for AI agents","github_url":"https://github.com/agi-inc/agent-protocol","owner":"agi-inc","repo":"agent-protocol","owner_avatar_url":"https://avatars.githubusercontent.com/u/130614869?v=4","primary_language":"Python","stars":1458,"forks":189,"topics":["agents","ai","ai-agent","api","auto-gpt","gpt-4","javascript","llms","openapi","protocol","python","typescript"],"archived":false,"github_pushed_at":"2025-04-08T06:10:21+00:00","maintenance_label":"Dormant","stars_delta_30d":4,"url":"https://www.graphcanon.com/tools/agi-inc-agent-protocol","markdown_url":"https://www.graphcanon.com/tools/agi-inc-agent-protocol.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/agi-inc-agent-protocol","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=agi-inc-agent-protocol","description":"Common interface for interacting with AI agents. The protocol is tech stack agnostic - you can use it with any framework for building agents.","homepage_url":"https://agentprotocol.ai","license":"MIT","open_issues":50,"watchers":17,"ai_summary":"Provides a standardized protocol enabling interaction with any AI agent, regardless of underlying tech stack.","readme_excerpt":"

\n \"agent\n

\n

\n\n \n \n \n \n \n \n \n

\n\nMaintained by **AGI, Inc.**\n## 📚 Docs\n\n**You can find more info in the [docs](https://agentprotocol.ai/).**\n\n## 🧾 Summary\n\nThe AI agent space is young. Most developers are building agents in their own way. This creates a challenge:\nIt's hard to communicate with different agents since the interface is often different every time.\nBecause we struggle with communicating with different agents, it's also hard to compare them easily.\nAdditionally, if we had a single communication interface with agents, it'd also make it easier developing devtools that works with agents out of the box.\n\nWe present the **Agent Protocol** - a single common interface for communicating with agents.\nAny agent developer can implement this protocol.\nThe Agent Protocol is an API specification - list of endpoints, which the agent\nshould expose with predefined response models.\nThe protocol is **tech stack agnostic**. Any agent can adopt this protocol no\nmatter what framework they're using (or not using).\n\nWe believe, this will help the ecosystem grow faster and simplify the integrations.\n\nWe're starting with a minimal core. We want to build upon that iteratively\nby learning from agent developers about what they actually need.\n\n## 🚀 The incentives to adopt the protocol\n\n- Ease with which you can use the benchmarks.\n- Other people can more easily use and integrate your agent\n- Enable building general devtools (for development, deployment and monitoring)\n that can be built on top of this protocol\n- You don’t need to write boilerplate API and you can focus on developing your\n agent\n\n## 🎯 Immediate goals of the protocol\n\nSet a general simple standard that would allow for easy to use benchmarking of\nagents. One of the primary goals of the protocol is great developer experience,\nand simple implementation on the end of agent developers. You just start your\nagent and that’s all you have to do.\n\n## 🗣️ Request for Comments\n\nIf you'd like to propose a change or an improvement to the protocol. Please\nfollow the [RFC template](./rfcs/template.md).\n\n## ⚙️ Components\n\n### [Protocol](./schemas/openapi.yml)\n\nThe most important part. It specifies which endpoints should the agent expose.\nThe protocol is defined in [OpenAPI specification](./schemas/openapi.yml).\n\n#### How does the protocol work?\n\nRight now the protocol is defined as a REST API (via the\n[OpenAPI spec](./schemas/openapi.yml)) with two essential routes for interaction with\nyour agent:\n\n- `POST /ap/v1/agent/tasks` for creating a new task for the agent (for example giving\n the agent an objective that you want to accomplish)\n- `POST /ap/v1/agent/tasks/{task_id}/steps` for executing one step of the defined task\n\nIt has also a few additional routes for listing the tasks, steps and downloading / uploading artifacts.\n\n### [SDK](https://github.com/AI-Engineer-Foundation/agent-protocol/tree/main/packages/sdk)\n\nThis is our implementation of the protocol. It’s a library that you can use to build your agent. You can use it, or you can implement it on your own. It’s up to you.\n\nUsing the SDK should simplify the implementation of the protocol to the bare minimum, but at\nthe same time it shouldn't tie your hands. The goal should be to allow agent\nbuilders to build their agents and the SDK should solve the rest.\n\nBasically it wraps your agent in a web server that allows for communication with\nyour agent (and in between agents in the future).\n\n### [Client](https://github.com/AI-Engineer-Foundation/agent-protocol/tree/main/packages/client)\n\nThis library should be u","github_created_at":"2023-07-06T14:39:13+00:00","created_at":"2026-07-11T23:46:45.825136+00:00","updated_at":"2026-10-08T06:00:43.647518+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"}],"tags":[{"slug":"agents","name":"agents"},{"slug":"ai-agent","name":"ai-agent"},{"slug":"api","name":"api"},{"slug":"auto-gpt","name":"auto-gpt"},{"slug":"gpt-4","name":"gpt-4"},{"slug":"javascript","name":"javascript"},{"slug":"llms","name":"llms"},{"slug":"openapi","name":"openapi"}],"trust":{"provenance":{"is_fork":false,"github_id":663095187,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-10-08T06:00:42.794Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":547,"last_release_at":"2024-04-02T18:06:10Z","stars_delta_30d":4,"open_issues_delta_30d":1},"security_summary":{"status":"findings","scanner":"osv@v1","low_count":35,"high_count":0,"last_scan_at":"2026-07-11T23:46:47.238Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"mcp":{"source":"repo_scan","observed_at":"2026-10-08T06:00:43.326Z","server_manifest":false},"scan":{"source":"repo_scan","observed_at":"2026-10-08T06:00:43.326Z"},"languages":{"value":["python","javascript","typescript"],"source":"github.language+package.json","observed_at":"2026-10-08T06:00:43.326Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-10-08T06:00:43.326Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When you want to ensure interoperability between different AI agents irrespective of the frameworks used by them.","To facilitate easier integration of your AI agent into a broader ecosystem that supports the Agent Protocol for seamless communication."],"when_not_to_use":["If you are developing an isolated system with no intention to communicate or integrate with other AI agents outside this scope.","When working in environments where specific, proprietary interfaces provide significantly better performance or features than adhering to a generic protocol could offer."],"source":"enrich:decision_facts","observed_at":"2026-07-16T20:50:20.390Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"agent-protocol provides a standardized interface for interacting with various AI agents regardless of their underlying tech stack, aiming to ease development, deployment, and benchmarking."}]}}