JSON Agents defines the Portable Agent Manifest (PAM) — a universal, JSON-native format for describing AI agents, their capabilities, tools, runtimes, and governance in a single, interoperable manifest.

It enables seamless agent portability across frameworks, platforms, and ecosystems without code translation.

• Framework Interoperability: Convert between LangChain, OpenAI, AutoGen, MCP, and custom frameworks
• Universal Standard: One manifest format that works everywhere
• Built-in Governance: Security policies, sandboxing, and audit trails included
• Multi-Agent Systems: Orchestrate complex agent graphs with conditional routing
• Production Ready: Complete with URI scheme (ajson://) and policy expression language
• Schema Validated: JSON Schema 2020-12 ensures correctness

{
  "manifest_version": "1.0",
  "agent": {
    "id": "ajson://example.com/agents/hello",
    "name": "Hello Agent",
    "version": "1.0.0"
  },
  "capabilities": [
    {
      "id": "generation",
      "description": "Generate friendly greetings"
    }
  ]
}

{
  "manifest_version": "1.0",
  "profiles": ["core", "exec", "gov", "graph"],
  "agent": {
    "id": "ajson://example.com/agents/support-hub",
    "name": "Customer Support Hub"
  },
  "capabilities": [
    { "id": "routing" },
    { "id": "qa" },
    { "id": "classification" }
  ],
  "runtime": {
    "type": "node",
    "entrypoint": "dist/hub.js"
  },
  "security": {
    "sandbox": "container"
  },
  "policies": [
    {
      "id": "deny-external",
      "effect": "deny",
      "action": "tool.call",
      "where": "tool.endpoint !~ 'internal.corp'"
    }
  ],
  "graph": {
    "nodes": [
      { "id": "router", "ref": "ajson://example.com/agents/router" },
      { "id": "faq", "ref": "ajson://example.com/agents/faq" },
      { "id": "escalation", "ref": "ajson://example.com/agents/escalation" }
    ],
    "edges": [
      { "from": "router", "to": "faq", "condition": "message.intent == 'faq'" },
      { "from": "router", "to": "escalation", "condition": "message.priority > 8" }
    ]
  }
}

7 standard capabilities with formal JSON schemas:

• Summarization — Condense content intelligently
• Routing — Direct messages by intent
• Retrieval — Query knowledge sources
• QA — Answer questions with context
• Classification — Categorize inputs
• Extraction — Extract structured data
• Generation — Create new content

Formal ajson:// URI scheme (RFC 3986 compliant):

ajson://example.com/agents/router
→ https://example.com/.well-known/agents/router.agents.json

Declarative access control with rich operators:

{
  "where": "tool.type == 'http' && tool.endpoint !~ 'external'"
}

Progressive enhancement with 4 profiles:

• Core — Identity, capabilities, tools (required)
• Exec — Runtime metadata (optional)
• Gov — Security and policies (optional)
• Graph — Multi-agent orchestration (optional)

• Framework Migration: Move agents between LangChain, OpenAI, AutoGen without rewriting
• Agent Catalogs: Build discoverable registries of reusable agents
• Orchestration: Define complex multi-agent workflows with conditional routing
• Enterprise Governance: Enforce security policies and maintain audit trails
• Marketplaces: Standardized format for distributing and monetizing agents
• CI/CD Testing: Schema-based validation for automated testing pipelines

We welcome contributions of all kinds:

• Bug Reports: Found an issue? Let us know!
• Feature Proposals: Have an idea? Open a discussion!
• Documentation: Help improve our docs
• Tools: Build validators, converters, or integrations
• Framework Support: Add mappings for new frameworks

See our Contributing Guide for details.

Code of Conduct: This project follows the Contributor Covenant 2.0.

• ✅ Core, Exec, Gov, Graph profiles
• ✅ 7 capability schemas with formal validation
• ✅ URI scheme specification (ajson://)
• ✅ Policy expression language
• ✅ Framework mapping guide
• ✅ Complete documentation

• Reference validator implementations (Python, Node.js, Go)
• Framework converter CLI tools
• Public registry service with API
• Additional capability schemas (dialogue, planning, reasoning)
• Community extension showcase

• Real-time profile for streaming agents
• Evaluation profile for testing/benchmarking
• Enhanced policy expression functions
• Formal IETF/W3C standardization track
• Enterprise support packages

JSON AGENTS is built on established foundations:

• ✅ RFC 8259 — JSON Data Interchange Format
• ✅ RFC 3986 — URI Generic Syntax
• ✅ RFC 2119 — Requirement Levels
• ✅ ECMA-404 — JSON Data Interchange Syntax
• ✅ ISO/IEC 21778:2017 — JSON Standard
• ✅ JSON Schema 2020-12 — Validation

JSON AGENTS is released under the Apache 2.0 License.

See LICENSE for details.

JSON agents draws inspiration from:

• JSON Schema and JSON-LD communities
• OpenAPI and AsyncAPI specifications
• Agent framework developers (LangChain, AutoGen, CrewAI)
• Model Context Protocol contributors
• JsonResume.org
• The broader open-source AI community

Special thanks to all contributors who have helped shape this standard.