Skip to content

getting-started.md

Latest commit

 

History

History
99 lines (66 loc) · 2.6 KB

getting-started.md

File metadata and controls

99 lines (66 loc) · 2.6 KB

LiAgent OS Getting Started

This is the shortest path to a first successful run of LiAgent OS.

Recommended starting point: API bootstrap mode. It lets you confirm that the runtime works before you invest time in local models, model paths, TTS/STT assets, or hybrid tuning.

What you need

  • Python 3.10+
  • one compatible LLM API key
  • about 5 minutes

Success target

By the end of this guide, you should be able to:

  • start the Web UI on http://127.0.0.1:8080
  • send a prompt and receive a response
  • see runtime events in the interface

1. Install the project

cd /path/to/liagent_git
python3 -m venv venv
./venv/bin/pip install -e .

Optional extras can wait until later:

./venv/bin/pip install -e '.[browser]'
./venv/bin/pip install -e '.[discord]'
./venv/bin/pip install -e '.[mcp]'

2. Create local config files

cp config.example.json config.json
cp .env.example .env

3. Add only the minimum API settings

Put your API key in .env:

LLM_API_KEY=your_api_key

Optional overrides:

# LLM_API_BASE_URL=https://api.openai.com/v1
# LLM_API_MODEL=gpt-4o

You do not need to configure local model paths for this first run.

4. Start the Web UI

./venv/bin/liagent --web --host 127.0.0.1 --port 8080

Then open http://127.0.0.1:8080.

5. Confirm the first successful run

You should now see:

  • the Web UI loading on localhost
  • a response when you send a simple message
  • runtime events, not just a plain chat bubble

Try one of these first prompts:

  • Summarize the AI agent releases from this week that are worth tracking.
  • Break this request into an execution plan, research first, then give me the final proposal.
  • Ask me before any file writes, command execution, or higher-risk actions.

6. If it does not work

Check these first:

  • Missing response: verify LLM_API_KEY is set in .env
  • Startup path validation errors: keep the first run in API bootstrap mode and do not add local model paths yet
  • Browser/tooling features missing: install the .[browser] extra and finish the Playwright setup later
  • Discord confirmations missing: add the .[discord] extra and verify the required Discord environment variables

7. What to do next

After the first successful run: