REAL SCENARIOS

Use cases: reusable sign-in and file retrieval

Two recorded uses: a person signs in once and a later run reuses saved state; an Agent reads official Python docs and retrieves a verified 1,875-byte file.

These two tasks were recorded in October 2026 on service version 0.1.12. Site rules, product versions, and charges may change.

01 / REUSE A SIGN-IN

You sign in now.
A later task can reuse it.

In a recorded Feishu task, a person completed the first sign-in and the browser state was saved at the end of that run. A new run using the same saved state opened the signed-in message workspace without entering credentials again.

  1. Complete the first sign-in and save state.
  2. Load that state into a later run.
  3. Check the actual page to verify the sign-in still works.

The two runs cost 5.84 points at the time. Check current pricing and the final charge for your own task. The target site decides how long a sign-in remains valid. The sign-in screen is withheld for privacy.

Read the handoff guide →
First taskHuman sign-in↓ Save stateLater taskOpen signed-in workspaceWorkflow illustration from the October 2026 record
02 / PUBLIC SOURCE TO FILE

Read the source.
Bring the file home.

An Agent opened the official Python 3.13 release notes, prepared a Markdown summary, and triggered a download in the cloud browser. It retrieved 1,875 bytes and verified both file length and SHA-256.

  1. Use a CDP client to read the public page.
  2. Prepare content and trigger a cloud download.
  3. Retrieve and verify the artifact before ending the run.

The Agent wrote the summary; Baizhi Cloud Browser supplied page access and file delivery. That run cost 1.13 points at the time; check current pricing and your final charge.

Read the file and lifecycle guide →
Real screenshot of the official Python 3.13 page visited in the recorded task
Official Python documentation visited in the October 2026 task.
About these cases

These tasks were recorded in October 2026, with only public visual content shown. The live setup reference is authoritative for API behavior.