Flipside CLI

Query blockchain data, create AI agents, and build automated data pipelines.

First Steps

Always start by checking what's available in the user's organization:

# 1. Verify authentication and see the current org
flipside whoami

# 2. List available agents (names vary by org)
flipside agents list

# 3. List available skills
flipside skills list

Agent and skill names are organization-specific. Always run flipside agents list to discover what agents are available before trying to use them.

Golden Rule

Use Flipside agents for SQL queries. Don't write SQL from scratch—let the agents generate correct queries against Flipside's data warehouse.

# One-shot query (no session memory)
flipside agents run <org>/<agent_name> --message "Get the top 10 DEX swaps on Ethereum today"

# Multi-turn conversation (maintains context between messages)
flipside chat create  # get session ID
flipside chat send-message --session-id <id> "Get top DEX swaps on Ethereum"
flipside chat send-message --session-id <id> "Now filter to only Uniswap"

# Always use --title when testing (makes sessions easier to find)
flipside agents run <org>/<agent_name> --message "test query" --title "Testing from Claude Code"

Primary Workflows

One-Shot Queries (agents run)

For single queries without session memory, use flipside agents run:

# First, list available agents to find the right one
flipside agents list

# Run a chat agent with a message
flipside agents run <org>/<agent_name> --message "Show me the largest ETH transfers today"

# Run a sub agent with structured JSON input
flipside agents run <org>/<agent_name> --data-json '{"chain": "ethereum", "limit": 10}'

# Add a title for easier session tracking
flipside agents run <org>/<agent_name> --message "Query request" --title "My analysis"

IMPORTANT: Agent names MUST use the org/agent_name format. Get the exact names from flipside agents list.

Multi-Turn Conversations (chat)

For conversations that need session memory (follow-up questions, context from previous messages), use chat:

# Create a new chat session
flipside chat create
# Returns: Session ID: abc123...

# Send messages to the session (maintains context)
flipside chat send-message --session-id <session-id> "What were the top DEX swaps on Ethereum yesterday?"

# Follow up (the session remembers previous context)
flipside chat send-message --session-id <session-id> "Now show me the same for Arbitrum"

# List your sessions
flipside chat list

# Resume an existing session interactively
flipside chat resume

Use chat when you need:

• Follow-up questions that reference previous answers
• Multi-step analysis where context matters
• Building on previous query results

Running SQL Directly

For known queries, you can run SQL directly:

# Create and execute a query
flipside query create "SELECT * FROM ethereum.core.fact_blocks LIMIT 10"

# Execute an existing saved query
flipside query execute <query-id>

# Get results from a query run
flipside query-run result <run-id>

Command Reference

Agents

flipside agents list                              # List available agents
flipside agents init my_agent                     # Create agent YAML template
flipside agents validate my_agent.agent.yaml      # Validate before deploy
flipside agents push my_agent.agent.yaml          # Deploy agent
flipside agents run <org>/<agent> --message "..." # Run chat agent
flipside agents run <org>/<agent> --data-json '{}' # Run sub agent
flipside agents describe <org>/<agent>            # Show agent details
flipside agents pull <org>/<agent>                # Download agent YAML
flipside agents delete <org>/<agent>              # Delete an agent

Chat

flipside chat                                     # Start interactive chat
flipside chat --run <run-id>                      # Chat about an automation run
flipside chat resume                              # Resume previous chat
flipside chat list                                # List chat sessions
flipside chat send-message --session-id <id> "x" # Programmatic message

Queries

flipside query create "SELECT ..."                # Create saved query
flipside query list                               # List saved queries
flipside query get <query-id>                     # Get query details
flipside query execute <query-id>                 # Execute query (creates run)
flipside query runs <query-id>                    # List runs for a query
flipside query-run result <run-id>                # Get query results
flipside query-run status <run-id>                # Check run status
flipside query-run poll <run-id>                  # Poll until complete

Automations

flipside automations list                         # List automations
flipside automations init my_pipeline             # Create automation YAML
flipside automations validate <file>     

...