// REFERENCE
Overview
A quick reference as to what is what with the http4k AI modules.
Universal LLM adapters#
| Provider | Chat | Streaming Chat | Image Generation | In-Memory Fake |
|---|---|---|---|---|
| AnthropicAI | ✅ | ❌ | ❌ | http4k-connect-ai-anthropic-fake |
| Azure | ✅ | ✅ | ❌ | http4k-connect-ai-openai-fake |
| Gemini | ✅ | ✅ | ❌ | http4k-connect-ai-openai-fake |
| Github Models | ✅ | ✅ | ❌ | http4k-connect-ai-openai-fake |
| Open AI | ✅ | ✅ | ✅ | http4k-connect-ai-openai-fake |
LangChain4J#
Plug-in http4k clients into any Langchain-compatible AI model, embedding, or vector store.
Model Context Protocol (Pro-tier)#
Type-safe implementation of the MCP protocol, over the HTTP Streaming, SSE and Standard IO transports. See the MCP SDK reference for details.
- MCP-SDK: for building Model Context Protocol (MCP) servers
- MCP-client: for connecting to Model Context Protocol (MCP) servers
- MCP-desktop-client: native desktop client to bridge MCP servers with Desktop clients (eg. Claude)
- MCP-x402: for x402 payment-protected MCP tools
- MCP-MPP: for Machine Payments Protocol payment-protected MCP tools
Agent2Agent (Pro-tier)#
Type-safe implementation of the A2A protocol, over both the JSON-RPC and REST/HTTP bindings. See the A2A SDK reference for details.
- A2A-SDK: for building A2A agents, including Agent Cards, Tasks, Streaming and Push Notifications
- A2A-client: for connecting to A2A agents
- MCP-A2A-bridge: for exposing an A2A agent to MCP clients as tools
Low-level Model API clients#
| Vendor | System | In-Memory Fake | Notes |
|---|---|---|---|
| AnthropicAI | API | ✅ | Includes content generators |
| AzureAI | API | ✅ | Includes content generators and GitHubModels compatibility |
| LM Studio | API | ✅ | |
| Ollama | API | ✅ | Includes content generators and image generation |
| Open AI | API | ✅ | Includes content generators and image generation |
| TypeSafe | API | ✅ | Typed question/answer API instead of prompts |
