How Browseflow works

1. Create a workflow

Sign up, create or join an organization, then open the app home. Start from a template or a blank canvas. Free orgs get 3 workflows; Pro is unlimited.

2. Add nodes

• Start — run manually from the Run button.
• Schedule — hourly or daily (save the graph so the scheduler can load it).
• Open URL, Act, Extract, Observe, Agent — browser steps powered by Stagehand + Browserbase.
• HTTP, Wait, Condition, Send Email — APIs, delays, gates, and notifications.

3. Wire data between steps

Click a node, focus a field, then click a Connection chip to insert tokens like {{nodeId.path}}. Upstream outputs are interpolated before each step runs.

4. Run, stop, replay

Hit Run to validate the graph, persist it, and start a background task. Watch live status in the console. After browser steps finish, open the session replay. Use Save graph from the ⋯ menu so scheduled workflows keep the latest snapshot.

Local dev: you must run npm run trigger:dev in a second terminal. If you only run Next.js, runs stay QUEUED then become EXPIRED.

Duplicate, export, import

In the editor ⋯ menu: Duplicate clones the graph with new node ids, Export JSON downloads a .browseflow.json file. On the app home, use Import to load that file into a new workflow.

Plans & usage

• Free — up to 3 workflows, 50 runs/month, all nodes except Agent. Usage meters live in the sidebar.
• Pro — unlimited workflows + Agent node.

Email delivery

Set RESEND_FROM_EMAIL to a verified Resend sender (e.g. Browseflow <alerts@yourdomain.com>). Without it, the app falls back to Resend's onboarding address for development.