Tapicker Bridge Overview
Tapicker Bridge lets you use Tapicker from a terminal, an application, or an AI agent. You can find and run your Workflows, create or update Workflows, check execution progress, and retrieve extracted data.
Your browser and the Tapicker Extension perform the automation. Bridge connects external callers to that Extension, so the browser must stay open and the Extension must be connected while a Workflow runs.
How It Fits Together
Terminal (CLI) ──────┐
Application (HTTP) ──┼── Tapicker Bridge ── Tapicker Extension ── Website
AI agent (MCP) ──────┘
For your first setup, run Bridge on the same computer as your browser. Start with Getting Started: install the CLI, connect the Extension, run a Workflow, and add Tapicker to your agent.
Choose an Interface
All three interfaces use the same Workflow and data capabilities. Choose based on how you want to call them:
| Interface | Use it for | Example |
|---|---|---|
| CLI | Terminal commands and shell scripts | tapicker workflow start WORKFLOW_ID --watch |
| HTTP | Applications and services | POST /v1/commands/workflow.start |
| MCP | AI agents that discover and call tools | workflow_start |
Bridge also provides status events so callers can wait for a Workflow to finish. Starting a Workflow and completing it are separate steps: a start response acknowledges startup; the final status tells you whether execution succeeded.
What You Can Do
- Use your Workflows: list, read, start, pause, resume, and stop them.
- Author Workflows: read the connected Extension's guides and schemas, validate node fragments, then create or update your Workflows.
- Inspect execution: read status and query the latest run's logs when needed.
- Access results: retrieve a data table or delete it from Tapicker Data.
Access uses the connected Extension's account, permissions, and plan limits. Reading all table rows consumes export quota. Generated Workflows still need to be checked against the actual website; static validation checks configuration, not whether a selector works on a live page.
Read in Order
- Getting Started: complete a local setup and your first run.
- CLI: use terminal commands and work with Workflow JSON.
- HTTP: make requests from your own application.
- MCP: configure an agent and guide it through Workflow creation.
- Remote Deployment: connect across computers with TLS and authentication.