Quickstart¶
Five minutes from pip install to a working agent running on a local model.
Package name vs import name
You install maki-framework but import maki.
1. Install Ollama and pull a small model¶
Install Ollama and start it (the desktop app starts the server for you; otherwise run ollama serve). Then pull a small model:
ollama pull llama3.2
Maki talks to Ollama at http://localhost:11434 by default. See Configuration to point it elsewhere.
2. Install Maki¶
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install maki-framework
Requires Python 3.10 or newer.
3. Ask the model a question¶
from maki import MakiLLama
llm = MakiLLama(model="llama3.2")
response = llm.chat("Explain recursion in one sentence.")
print(response.content)
4. Run your first agent¶
from maki import MakiLLama
from maki.agents import Agent
llm = MakiLLama(model="llama3.2")
agent = Agent(
name="Reviewer",
maki_instance=llm,
role="code reviewer",
instructions="Focus on bugs, regressions, and missing validation.",
)
print(agent.execute_task("Review this design: a plugin system with file access."))
5. Swap the backend¶
The same agent runs on a hosted model by changing one object:
from maki import MakiAnthropic # pip install "maki-framework[anthropic]"
llm = MakiAnthropic(model="claude-sonnet-4-5")
Troubleshooting¶
| Symptom | Fix |
|---|---|
Connection error to localhost:11434 |
Ollama is not running. Start the app or run ollama serve. |
model not found |
Run ollama pull <model> for the exact name you passed. |
| Request times out on a long answer | Use Agent(use_streaming=True); the timeout then applies per chunk. See Agents. |
ImportError mentioning maki[...] |
Install the extra it names, e.g. pip install "maki-framework[openai]". |