atmosphere
Portable AI agent runtime for the JVM. One @Agent class runs on Spring AI, LangChain4j, Anthropic, or 9 more behind one SPI. Token streaming, tool calls, human approvals, and governance over WebSocket, SSE, gRPC, or WebTransport/HTTP3. Speaks MCP, A2A, and AG-UI.
- stars
- 3,812
- 30-day movement
- +21/day
- Related entries
- 66
- Connections
- 3
Atmosphere is a JVM runtime for AI agents where a single @Agent class can run on Spring AI, LangChain4j, Anthropic, or nine other backends behind one SPI. It provides token streaming, tool calls, human approvals, and governance over WebSocket, SSE, gRPC, or WebTransport/HTTP3, and speaks MCP, A2A, and AG-UI.
Reach for it when you want to write agent code once on the JVM and swap LLM frameworks or transports without rewriting the agent.
Use it to
- Define agents with @Agent and switch backends via the SPI
- Stream tokens and tool calls over WebSocket, SSE, gRPC, or HTTP3
- Add human approval and governance steps to agent workflows
- Expose or consume agents through MCP, A2A, or AG-UI
- Run agents in Spring Boot or Quarkus services
For JVM developers building portable, governed AI agents
- Role
- mcp-server
- Language
- Java
- Licence
- Apache-2.0
- Forks
- 761
- Open issues
- 2
- Last push
- 2026-09-14
- Latest release
- atmosphere-4.0.0 · 2026-02-20
- Skills shipped
- 6