Agent Discovery

This is an Alpha version still in active development; its structure, interfaces, and functionality are changing frequently as the design evolves. Join the GitHub discussions to help shape it.

💬 Discussion: Agent metadata
💬 Discussion: Manifest-Based Agent Offline Discoverability

Agents can be discovered via the GET /agents endpoint. Discovery relies on agent details composed of:

Offline description created by the agent developer (in the form of a manifest).
Online description assembled at runtime by some underlying platform.

Agent Detail Model

name

string

required

Name of the agent

Example:

"chat"

description

string

required

Short description of the agent. This field might be consumed and interpreted by large language models (LLMs).

Example:

"Conversational agent with memory, supporting real-time search, Wikipedia lookups, and weather updates through integrated tools"

interfaces

object

required

Description of agent interfaces

metadata

object

Optional data to support agent discovery and cataloguing.

metadata.documentation

string

Full agent documentation in markdown.

Example:

"The agent is a conversational system designed to process user messages, maintain context, and generate..."

metadata.license

string

SPDX license ID: https://spdx.org/licenses/

Example:

"Apache-2.0"

metadata.programmingLanguage

string

Programming language, supported values: https://docs.github.com/en/get-started/learning-about-github/github-language-support

Example:

"Python"

metadata.naturalLanguages

string[]

List of human-readable languages supported by the agent for conversational interactions or content generation. Use ISO 639-1 codes.

metadata.framework

string

For example BeeAI, crewAI, Autogen, AG2

Example:

"BeeAI"

metadata.useCases

string[]

A list of short plain-text strings describing use cases supported by the agent. This field might be consumed and interpreted by large language models (LLMs).

metadata.examples

any[]

metadata.tags

(string | enum<string>)[]

metadata.createdAt

string

Timestamp in RFC3339 format (ISO8601-compatible).

metadata.updatedAt

string

Timestamp in RFC3339 format (ISO8601-compatible).

metadata.author

object

metadata.contributors

object[]

metadata.links

object[]

metadata.dependencies

object[]

Experimental: Represents a dependency required by the agent, specifying external resources or capabilities it relies upon, such as other agents, tools, or AI models.

metadata.recommendedModels

string[]

Model name (ollama, OpenAI)

status

object

Runtime metrics and statistics provided by the platform.

Agent RunHolding persistent state in the agent conversation

{
  "name": "chat",
  "description": "Conversational agent with memory, supporting real-time search, Wikipedia lookups, and weather updates through integrated tools",
  "interfaces": {
    "input": "chat",
    "output": "chat",
    "awaits": [
      {
        "name": "<string>",
        "request": {},
        "response": {}
      }
    ]
  },
  "metadata": {
    "documentation": "The agent is a conversational system designed to process user messages, maintain context, and generate...",
    "license": "Apache-2.0",
    "programmingLanguage": "Python",
    "naturalLanguages": [
      "en"
    ],
    "framework": "BeeAI",
    "useCases": [
      "Conversational AI – Handles multi-turn conversations with memory."
    ],
    "examples": [
      "<any>"
    ],
    "tags": [
      "Chat"
    ],
    "createdAt": "2023-11-07T05:31:56Z",
    "updatedAt": "2023-11-07T05:31:56Z",
    "author": {
      "name": "John Smith",
      "email": "jsmith@example.com",
      "url": "https://example.com"
    },
    "contributors": [
      {
        "name": "John Smith",
        "email": "jsmith@example.com",
        "url": "https://example.com"
      }
    ],
    "links": [
      {
        "type": "source-code",
        "url": "https://github.com/i-am-bee/beeai.git"
      }
    ],
    "dependencies": [
      {
        "type": "tool",
        "name": "weather"
      }
    ],
    "recommendedModels": [
      "llama3.3:70b-instruct-fp16"
    ]
  },
  "status": {
    "avgRunTokens": 123,
    "avgRunTimeSeconds": 123,
    "successRate": 50
  }
}

Introduction

How To

Core Concepts

Agents

Agent Communication Protocol

Observability & Telemetry

Agent Discovery

Agent Detail Model

Introduction

How To

Core Concepts

Agents

Agent Communication Protocol

Observability & Telemetry

​Agent Detail Model

Agent Detail Model