How to QA Your App with Computer Use with Eigent

How it works

1. Tell Eigent which app environment to test and which user flows matter most.
2. Eigent clicks through the flows, types into fields, and records what breaks.
3. For every bug found, Eigent documents repro steps, expected result, actual result, and severity.
4. Eigent continues past non-blocking issues and ends with a short triage summary.
5. Optionally ask Eigent to fix a found bug, create Linear issues, or narrow the next pass to one failing flow.

More prompts to try

• Test my app. Find any major issues and give me a report.
• Test signup, invite a teammate, and upgrade billing in staging. Log every bug with repro steps, expected result, actual result, and severity.
• Continue the QA pass and focus only on the checkout flow — I want to understand all the edge cases.
• Turn the P1 bugs from this report into draft Linear issues.

How to use

Tell Eigent what environment to test (local, staging, or production-like) and which flows to cover. Name the issue types you care about — broken functionality, layout issues, confusing copy, or all of the above. If account state, test data, or feature flags affect the flow, include that up front. If one blocking issue should end the run, say so; otherwise Eigent collects all issues and summarizes at the end. After the pass, ask Eigent to fix bugs, create tickets, or run a follow-up pass on a specific failing flow.

Expected output

A structured bug report listing every found issue with repro steps, expected vs. actual result, and severity. Ends with a short triage summary of P0–P2 findings and recommended next actions.

Limitations

• Computer Use exercises the UI directly — it cannot test server-side logic or background jobs not visible in the interface.
• Test accuracy depends on providing clear account state and environment details up front.
• Very long or complex test plans benefit from being split into focused passes rather than one long run.