Pal

A personal agent that learns your preferences, context, and history. Built with Agno.

Pal creates SQL tables on demand, remembers everything you tell it, and gets better over time using LearningMachine. It connects notes, people, bookmarks, and anything else you want to track.

What's Included

Agent	Pattern	Description
Pal	SQL + Exa MCP + LearningMachine	Personal knowledge system that dynamically creates schemas, searches the web, and learns your preferences.

Get Started

# Clone the repo
cp example.env .env
# Edit .env and add your OPENAI_API_KEY and EXA_API_KEY

# Start the application
docker compose up -d --build

Confirm Pal is running at http://localhost:8000/docs.

Connect to the Web UI

1. Open os.agno.com and login
2. Add OS → Local → http://localhost:8000
3. Click "Connect"

Try it:

Save a note: Met with Sarah Chen from Acme Corp
What do I know about Sarah?
Create a projects table and track my current projects

Deploy to Railway

Requires:

• Railway CLI
• OPENAI_API_KEY set in your environment

railway login

./scripts/railway_up.sh

The script provisions PostgreSQL, configures environment variables, and deploys your application.

Connect to the Web UI

1. Open os.agno.com
2. Click "Add OS" → "Live"
3. Enter your Railway domain

Manage deployment

railway logs --service pal      # View logs
railway open                    # Open dashboard
railway up --service pal -d     # Update after changes

To stop services:

railway down --service pal
railway down --service pgvector

How Pal Works

SQL Database — Pal's structured storage for your data. Notes, bookmarks, people, projects, decisions — tables are created on demand with a pal_ prefix. Uses SQLTools against the shared PostgreSQL database.

LearningMachine — Meta-knowledge about you and the database. Preferences, patterns, schemas created, query patterns that work. Stored in a separate vector knowledge base (pal_learnings).

Exa MCP — Web research via the Exa search API. Pal can look things up and optionally save findings to the database.

Project Structure

├── pal/
│   └── agent.py              # Pal agent implementation
├── app/
│   ├── main.py               # AgentOS entry point
│   └── config.yaml           # Quick prompts config
├── db/
│   ├── session.py            # PostgreSQL database helpers
│   └── url.py                # Connection URL builder
├── scripts/                  # Helper scripts
├── compose.yaml              # Docker Compose config
├── Dockerfile
└── pyproject.toml            # Dependencies

Common Tasks

Add tools to Pal

Agno includes 100+ tool integrations. See the full list.

from agno.tools.slack import SlackTools

pal = Agent(
    ...
    tools=[
        SQLTools(db_url=db_url),
        MCPTools(url=EXA_MCP_URL),
        SlackTools(),
    ],
)

Add dependencies

1. Edit pyproject.toml
2. Regenerate requirements: ./scripts/generate_requirements.sh
3. Rebuild: docker compose up -d --build

Use a different model provider

1. Add your API key to .env (e.g., ANTHROPIC_API_KEY)
2. Update pal/agent.py to use the new provider:

from agno.models.anthropic import Claude

model=Claude(id="claude-sonnet-4-5")

3. Add dependency: anthropic in pyproject.toml

---

Local Development

For development without Docker:

# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh

# Setup environment
./scripts/venv_setup.sh
source .venv/bin/activate

# Start PostgreSQL (required)
docker compose up -d pal-db

# Run the app
python -m app.main

Environment Variables

Variable	Required	Default	Description
OPENAI_API_KEY	Yes	-	OpenAI API key
EXA_API_KEY	No	-	Exa API key for web research
PORT	No	8000	API server port
DB_HOST	No	localhost	Database host
DB_PORT	No	5432	Database port
DB_USER	No	ai	Database user
DB_PASS	No	ai	Database password
DB_DATABASE	No	ai	Database name
RUNTIME_ENV	No	prd	Set to dev for auto-reload

Learn More

• Agno Documentation
• AgentOS Documentation
• Agno Discord