# Agent Resource Graph

> Source-backed static graph for agent frameworks, MCP servers, hosts, runtime environments, compatibility facts, comparisons, scenarios, and FAQ entries.

Schema version: 2026-05-18.1

## Graph summary
Total entities: 38
  agent-frameworks: 15
  mcp-servers: 10
  agent-hosts: 8
  agent-runtimes: 5
  agent-capabilities: see /agent-capabilities.json
Total relationships: 259
Compatibility facts: 52
Comparisons: 10
Scenarios: 8
FAQ entries: 19

## Relationships by type
- supports: 127
- compatible_with: 52
- works_with: 41
- best_for: 34
- exposes: 5

## Compatibility by type
- runtime_deployment: 22
- server_host: 20
- tool_permission: 5
- framework_mcp: 4
- server_transport: 1

## Entities
- AutoGen (agent-framework): Microsoft open-source framework for multi-agent conversation patterns, agent teams, and collaborative task solving.
- Cloudflare Agents SDK (agent-framework): Cloudflare framework for building and deploying stateful AI agents on Workers and Durable Objects.
- Coze (agent-framework): AI bot development platform by ByteDance with visual workflow builder, plugin ecosystem, and multi-model orchestration.
- CrewAI (agent-framework): Multi-agent automation framework for coordinating role-based agents, tasks, tools, and workflows.
- Dify (agent-framework): Open-source LLM application development platform with visual workflow builder, RAG pipeline, and multi-model support.
- Google Agent Development Kit (agent-framework): Open-source framework for building, evaluating, and deploying AI agents with multi-agent workflow support.
- Haystack (agent-framework): Open-source AI orchestration framework for pipelines, agents, retrieval, tools, and production RAG applications.
- LangChain / LangGraph (agent-framework): Agent and graph runtime ecosystem for building tool-using, stateful, and observable LLM applications.
- LlamaIndex (agent-framework): Data-oriented agent and workflow framework for building LLM agents over private data, tools, retrieval, and workflows.
- Mastra (agent-framework): TypeScript agent framework for building AI applications with agents, tools, workflows, and application integrations.
- Microsoft Agent Framework (agent-framework): Microsoft framework for building, orchestrating, and deploying agents and multi-agent workflows across Python and .NET.
- OpenAI Agents SDK (agent-framework): Framework for building agents with tools, handoffs, guardrails, tracing, and model orchestration.
- Pydantic AI (agent-framework): Python agent framework from the Pydantic team focused on type-safe agent development and structured outputs.
- Semantic Kernel (agent-framework): Microsoft SDK for integrating AI orchestration, plugins, agents, and enterprise application workflows.
- Vercel AI SDK (agent-framework): TypeScript toolkit for building AI-powered applications, streaming interfaces, tools, and agent-like workflows.
- Brave Search MCP Server (mcp-server): MCP server for web search and content discovery via Brave Search API, providing real-time results to agents.
- Cloudflare MCP Server (mcp-server): MCP server for connecting agent workflows to Cloudflare API and developer platform context.
- Context7 MCP Server (mcp-server): MCP server for retrieving current documentation and code examples for libraries and frameworks.
- Filesystem MCP Server (mcp-server): MCP server for controlled local filesystem access, including reading and writing files within configured directories.
- GitHub MCP Server (mcp-server): MCP server entry for GitHub repository, issue, pull request, and coding workflow context.
- Hugging Face MCP Server (mcp-server): MCP server entry for Hugging Face Hub model, dataset, and agent integration workflows.
- Notion MCP Server (mcp-server): MCP server for querying and managing Notion workspaces including databases, pages, and search.
- Playwright MCP Server (mcp-server): MCP server from Microsoft for browser automation and testing workflows backed by Playwright.
- PostgreSQL MCP Server (mcp-server): MCP server category for exposing PostgreSQL schema and query context to agent workflows.
- Sentry MCP Server (mcp-server): MCP server for exposing Sentry errors, issues, and observability context to coding assistants.
- Claude Code (agent-host): Terminal coding agent host that can connect to external tools through MCP and project-level configuration.
- Claude Desktop (agent-host): Desktop host for Claude that can connect to local MCP servers and reviewed desktop extensions.
- Codex (agent-host): Coding agent host that can consume structured project context, tools, and repository instructions.
- Cursor (agent-host): AI code editor and agent host environment with support for connecting tools and project context.
- GitHub Copilot (agent-host): Coding assistant and host environment for repository-aware development workflows.
- Visual Studio Code (agent-host): IDE host for GitHub Copilot and MCP server configuration in coding-agent workflows.
- Windsurf (agent-host): AI-native IDE by Codeium with deep agent integration for code generation, refactoring, and contextual workflows.
- Zed (agent-host): IDE host with AI features and MCP server integration for editor-based agent workflows.
- Browser Agent Runtime (agent-runtime): Browser-based execution environment for agent workflows, supporting sandboxed UI, local storage, and Web API orchestration.
- Edge Worker Runtime (agent-runtime): Edge compute runtime for deploying agent workflows near users with serverless isolation.
- Local Desktop Runtime (agent-runtime): Local machine runtime for desktop-hosted MCP servers, filesystem tools, browser automation, and IDE agent workflows.
- Serverless / Container Runtime (agent-runtime): Deployment target family for stateless agents, tool gateways, MCP servers, and isolated execution.
- Workflow Orchestration Runtime (agent-runtime): Runtime category for durable, multi-step agent workflows with retries, routing, and state transitions.

## Scenarios
- Best agent frameworks for browser automation agents: /best-agent-frameworks-for-browser-automation-agent
- Best agent frameworks for coding agents: /best-agent-frameworks-for-coding-agent
- Best agent frameworks for data analysis agents: /best-agent-frameworks-for-data-analysis-agent
- Best agent frameworks for edge agent workflows: /best-agent-frameworks-for-edge-agent-workflow
- Best agent frameworks for enterprise agent workflows: /best-agent-frameworks-for-enterprise-agent-workflow
- Best agent frameworks for frontend agent UIs: /best-agent-frameworks-for-frontend-agent-ui
- Best agent frameworks for RAG workflows: /best-agent-frameworks-for-rag-workflow
- Best agent frameworks for research agents: /best-agent-frameworks-for-research-agent

## Comparisons
- AutoGen vs CrewAI: /compare/autogen-vs-crewai
- Cloudflare Agents SDK vs Vercel AI SDK: /compare/cloudflare-agents-sdk-vs-vercel-ai-sdk
- Context7 MCP Server vs Filesystem MCP Server: /compare/context7-mcp-server-vs-filesystem-mcp-server
- LangChain / LangGraph vs LlamaIndex: /compare/langchain-langgraph-vs-llamaindex
- LlamaIndex vs Haystack: /compare/llamaindex-vs-haystack
- Microsoft Agent Framework vs OpenAI Agents SDK: /compare/microsoft-agent-framework-vs-openai-agents-sdk
- OpenAI Agents SDK vs LangChain / LangGraph: /compare/openai-agents-sdk-vs-langchain-langgraph
- Semantic Kernel vs Microsoft Agent Framework: /compare/semantic-kernel-vs-microsoft-agent-framework
- Sentry MCP Server vs GitHub MCP Server: /compare/sentry-mcp-server-vs-github-mcp-server
- Vercel AI SDK vs Mastra: /compare/vercel-ai-sdk-vs-mastra

## Machine-readable feeds
- /entities.json
- /relationships.json
- /relationship-summary.json
- /geo.json
- /agent-graph.json
- /mcp-servers.json
- /agent-hosts.json
- /agent-runtimes.json
- /compatibility-summary.json
- /compatibility-matrix.json
- /agent-capabilities.json
- /agent-tools.json
- /faq.json
- /source-verification.json
- /route-manifest.json
- /feed-contracts.json