Skip to content
Overmind
Get started
Guides
Getting Started

Getting Started

Install Overmind and run your first optimization in about 10 minutes.

Install, init, open your IDE, type three commands. That’s the whole workflow.

Who this is for: data-backed Python agents with a clear input/output and a notion of “correct.” See the Platform Overview for context and examples/ for runnable agents.

Requirements

  • Python 3.10 or higher
  • Cursor or Claude Code
  • API keys for at least one LLM provider (OpenAI, Anthropic)

1. Install Overmind

Section titled “1. Install Overmind”
Terminal window
pip install overmind

2. Initialize the project

Section titled “2. Initialize the project”

From your agent’s project root:

Terminal window
cd your-agent-project/
overmind init

This creates .overmind/, collects your API keys, sets your analyzer model, and installs the Agent Skills into your IDE. You can re-run it at any time.

3. Register, configure, and optimize

Section titled “3. Register, configure, and optimize”

Open your project in Cursor or Claude Code. Run these three commands in the chat panel, in order:

/overmind-register-agent path/to/your/agent.py
/overmind-generate-spec-and-dataset my-agent
/overmind-optimize-agent my-agent

Each skill reads your codebase, asks what it can’t infer, and handles the rest. Results are pushed to console.overmindlab.ai/agents as optimization runs.

See the How to Use Overmind guide for what each skill does and what to expect at each step.

Instrumenting a running app (optional)

Section titled “Instrumenting a running app (optional)”

If you want Overmind traces from a deployed application, independently of the optimization workflow, install the Python or JS/TS tracing SDK and call init() once at startup:

Terminal window
pip install overmind
import overmind
overmind.init(service_name="my-service", environment="production")

See the Python SDK reference.

Tracing is independent of optimization. Use one, both, or neither.

Next steps

Section titled “Next steps”