Chat is the simplest way to work with SignalPilot agents. Instead of writing code from scratch, you can describe what you want in plain language. The agent then responds with code, explanations, or edits directly inside your notebook.

How Chat Works

• Natural language in, code out – type your request in the chat panel.

• Add context: using @mention, add cells, datasets, database to your context.

• Inline actions – ask to edit, fix, or refactor existing cells.

• Context-aware – chat remembers your datasets, connections, and prior steps in the session.

• Interactive loop – you can accept, reject, or modify agent suggestions before execution.

Example Interactions

• “Load my CSV file and show me the first 5 rows.”

• “Connect to my Postgres database and get total sales by region.”

• “Fix the error in the last cell and explain what went wrong.”

• “Make the plot use a log scale on the y-axis.”

Chat turns your notebook into a conversation-driven workflow: you describe your goals, and the agent writes and edits the code to make it happen.

FAQ

Q: Can I edit cells through chat?
Yes. You can ask the agent to modify an existing cell, and it will suggest an inline edit.

Q: Does chat remember everything across sessions?
No. Context is remembered during your current session only. SignalPilot follows a zero data retention policy outside your notebook.

Q: What if the agent makes a mistake?
You can reject suggestions, undo changes, or refine your request for a better result.