Unmanaged Agent

Open sourced agent infra.

How It Works

1. Create an Agent

Define the model, system prompt, tools, MCP servers, and skills. Create the agent once and reference it by ID across sessions.

For the first version, only pi-agent is supported as the agent runtime. Support for Claude Agent SDK, OpenAI Agents SDK, and Google ADK is coming soon.

2. Create an Environment

Configure a cloud container with pre-installed packages (Python, Node.js, Go, etc.), network access rules, and mounted files.

Sandbox provided by E2B and Daytona. Support for Blaxel and Modal is coming soon.

3. Start a Session

Launch a session that references your agent and environment configuration.

4. Send Events and Stream Responses

Send user messages as events. The agent autonomously executes tools and streams back results via server-sent events (SSE). Event history is persisted server-side and can be fetched in full.

Quick Start

1. Create an agent

curl -X POST http://localhost:8000/agents \
  -H "Content-Type: application/json" \
  -d '{
    "name": "coding-assistant",
    "model": "claude-sonnet-4-6",
    "runtime": "pi-agent",
    "system_prompt": "You are a helpful coding assistant."
  }'

2. Create an environment

curl -X POST http://localhost:8000/environments \
  -H "Content-Type: application/json" \
  -d '{
    "name": "python-sandbox",
    "sandbox_provider": "e2b",
    "config": {"packages": ["python3", "nodejs"]}
  }'

3. Create a session

curl -X POST http://localhost:8000/sessions \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "<agent_id>",
    "environment_id": "<environment_id>",
    "title": "My first session"
  }'

4. Send a message and stream the response

curl -N -X POST http://localhost:8000/sessions/<session_id>/events \
  -H "Content-Type: application/json" \
  -d '{"type": "message", "content": {"text": "Write a Python script that prints the first 10 Fibonacci numbers."}}'

Self-Hosting Guide

Prerequisites

• Python 3.14+
• An E2B API key (for sandbox environments)
• At least one LLM API key (Anthropic, OpenAI, Gemini, or OpenRouter)

Option 1: Run directly

cd api-server

# Copy and fill in your API keys
cp .env.example .env

# Install dependencies
uv sync

# Start the server
uv run uvicorn main:app --host 0.0.0.0 --port 8000

Option 2: Docker

cd api-server

docker build -t unmanaged-agent .

docker run -p 8000:8000 \
  -e E2B_API_KEY=your_e2b_key \
  -e ANTHROPIC_API_KEY=your_anthropic_key \
  unmanaged-agent

The API docs are available at http://localhost:8000/docs.

Roadmap

• Metadata storage: Currently using SQLite. PostgreSQL/Neon DB support is on the way.
• SDKs: Python and TypeScript client SDKs.