Skip to main content
The chat REPL provides a conversational interface for data exploration with AI agents.

​
Starting a Chat

# Start a new interactive session
flipside chat

# Chat about a specific automation run
flipside chat --run <run-id>

# Resume an existing chat session
flipside chat resume

​
Interactive Mode

When you run flipside chat, you enter an interactive REPL where you can:
  • Ask questions in natural language
  • Get SQL queries generated and executed
  • Explore results and ask follow-up questions
  • Save and export your findings
Example session: 
> What were the top 10 NFT collections by volume yesterday?

I'll query the NFT sales data for you...

[Results displayed]

> Can you break that down by marketplace?

Sure, let me add marketplace grouping...

[Updated results]

​
Programmatic Chat Access

For scripting and automation: 
# Create a new chat session
flipside chat create

# Send a message and stream the response
flipside chat send-message "What were the top NFT sales yesterday?"

# Send a message and wait for complete response (non-streaming)
flipside chat send-message-sync "Summarize Ethereum gas usage this week"

​
Managing Chat Sessions

# List all your chat sessions
flipside chat list

# List messages in a specific session
flipside chat list-messages <session-id>

# List attachments in a session
flipside chat list-attachments <session-id>

# Resume a previous session
flipside chat resume

​
Chat Commands

CommandDescription
flipside chatStart interactive REPL
flipside chat createCreate a new session
flipside chat resumeResume existing session
flipside chat listList all chat sessions
flipside chat list-messages <id>List messages in a session
flipside chat send-message <msg>Send message (streaming)
flipside chat send-message-sync <msg>Send message (wait for complete response)
flipside chat list-attachments <id>List attachments in a session

​
Chat with Automation Context

Start a chat about a specific automation run to discuss results: 
flipside chat --run <automation-run-id>
This preloads the chat with context about the automation and its outputs, making it easy to ask questions about the results.