Quick Start¶

Get a Rune agent running in a few minutes.

1. Start the Daemon¶

The Rune daemon runs the gateway (HTTP API) and control plane.

Gateway: http://localhost:8080 — invoke agents, sessions, A2A
Control plane: http://localhost:8081 — deploy agents, manage replicas

rune daemon start --foreground

Leave this running in a terminal.

2. Set Your API Key¶

# Anthropic (Claude) — recommended
export ANTHROPIC_API_KEY=sk-ant-...

# Or OpenAI
export OPENAI_API_KEY=sk-...

# Or Gemini
export GEMINI_API_KEY=...

3. Deploy an Agent¶

Deploy the included basic example:

cd examples/basic-example
rune compose up -f rune-compose.yml

Or deploy any agent directory directly:

rune run ./examples/basic-example

4. Check Status¶

rune status

You should see the chat deployment and its replica health.

5. Invoke the Agent¶

Via CLI:

rune run --agent chat "Hello, what can you do?"

Via HTTP:

curl -X POST http://localhost:8080/v1/agents/chat/invoke \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello, what can you help me with?"}'

Streaming (SSE):

curl -X POST http://localhost:8080/v1/agents/chat/invoke \
  -H "Content-Type: application/json" \
  -d '{"input": "Tell me about yourself.", "stream": true}'

6. Multi-Turn Sessions¶

Create a session for persistent multi-turn conversations:

# Create a session
SESSION=$(curl -s -X POST http://localhost:8080/v1/agents/chat/sessions \
  -H "Content-Type: application/json" -d '{}' | jq -r .session_id)

# First turn
curl -X POST http://localhost:8080/v1/agents/chat/invoke \
  -H "Content-Type: application/json" \
  -d "{\"input\": \"My name is Alice.\", \"session_id\": \"$SESSION\"}"

# Second turn — agent remembers the full conversation history
curl -X POST http://localhost:8080/v1/agents/chat/invoke \
  -H "Content-Type: application/json" \
  -d "{\"input\": \"What is my name?\", \"session_id\": \"$SESSION\"}"

7. Stop and Clean Up¶

rune agent stop chat       # scale to 0 replicas
rune agent rm chat         # remove the deployment
rune daemon stop           # stop the daemon

Next Steps¶

First Agent — Create an agent from scratch with a custom tool
Examples — Browse all example agents (memory, web research, multi-agent, etc.)
Compose — Deploy multi-agent stacks
CLI Reference — All commands and options