HTTP-native event streaming for AI agents. Built in Go for performance and single-binary deployment.
- Single Binary: One executable, zero dependencies
- Fast: 10x faster than Python prototype
- Consumer Groups: Built-in load balancing with visibility timeouts
- SQLite Backend: Zero external dependencies, WAL mode for concurrency
- HTTP API: Simple REST + WebSocket
# Build and run
go build -o stream0 .
./stream0
# Server runs on http://127.0.0.1:8080curl -X POST http://localhost:8080/topics \
-H "Content-Type: application/json" \
-d '{"name": "tasks", "retention_days": 7}'curl -X POST http://localhost:8080/topics/tasks/messages \
-H "Content-Type: application/json" \
-d '{"payload": {"action": "analyze"}}'curl "http://localhost:8080/topics/tasks/messages?group=workers&max=10&timeout=5"curl -X POST http://localhost:8080/messages/{id}/ack \
-H "Content-Type: application/json" \
-d '{"group": "workers"}'python3 test_go.pyEnvironment variables:
STREAM0_SERVER_HOST- Bind address (default: 127.0.0.1)STREAM0_SERVER_PORT- Port (default: 8080)STREAM0_DB_PATH- Database path (default: ./stream0.db)
┌──────────┐ ┌──────────┐ ┌──────────┐
│ Agent A │────→│ stream0 │────→│ Agent B │
│ Producer │ │ (Go) │ │ Consumer │
└──────────┘ └────┬─────┘ └──────────┘
│
┌────┴────┐
│ SQLite │
│ (WAL) │
└─────────┘
- Single binary: ~20MB
- Startup time: <100ms
- Throughput: 10K+ messages/sec
- Memory: ~10MB baseline
MIT