Quickstart

import { Steps } from ‘@astrojs/starlight/components’;

1. Create a project

   tama init my-project
   cd my-project

   This creates:

   my-project/
   ├── tama.toml
   ├── agents/
   └── skills/

2. Add your first agent

   tama add oneshot summarizer

   This scaffolds agents/summarizer/AGENT.md:

   ---
   name: summarizer
   description: Summarizes text concisely.
   version: "1.0.0"
   pattern: oneshot
   call:
     model:
       role: thinker
   ---
   
   You are a concise summarizer. Given any text, return a clear 2-3 sentence summary.

3. Set your API key and model

   export ANTHROPIC_API_KEY=sk-ant-...
   export TAMA_MODEL_THINKER=anthropic:claude-sonnet-4-6

4. Run it

   tamar "The Rust programming language is a systems language that guarantees memory safety \
   without a garbage collector, using a borrow checker to enforce ownership rules at compile time."

   Output:

   Rust is a systems programming language that ensures memory safety at compile time through
   its ownership and borrow checker system, eliminating the need for a garbage collector.

5. Try a more powerful pattern

   Add a research agent with web search:

   tama add react researcher
   tama add skill search-web

   Edit agents/researcher/AGENT.md:

   ---
   name: researcher
   description: Researches topics using web search.
   version: "1.0.0"
   pattern: react
   call:
     model:
       role: thinker
     uses:
       - search-web
   ---
   
   You are a research assistant. Use search-web to find current information.
   When you have a comprehensive answer, call finish with key="done" and your findings.

   Point tama.toml at the new entrypoint:

   [project]
   name = "my-project"
   entrypoint = "researcher"

   Run:

   tama run "What are the latest developments in fusion energy?"

What just happened?

• tama init created the project scaffold
• tama add generated an agent with the correct file structure
• tama run executed the agent graph starting from the entrypoint
• The agent received your input as its first start() call and ran until finish()

Next steps

• Core Concepts — understand how data flows through agents
• Patterns overview — choose the right pattern for your use case
• Writing a Skill — give agents access to external tools