Continue with extract, quality checks, and export.
Tip: keep URLs on the same host for one run.
--
Ready to run Analyze.
Talk to Flowscrape AI
Ask questions like a real buyer. The assistant should explain value, answer concerns, and show why Flowscrape is worth buying.
OpenAI chat
Loading chat…
Workflow steps
Run core scraping first, then validate quality, then review/export results.
1) Core flow
Analyze page structure, get data snapshots, then extract products.
Troubleshooting tools (optional)
Analyze blocked? Paste or upload HTML to run analyze directly.
No file selected.
2) Results & export
Preview extracted products, review OpenAI’s sampled field suggestions, then confirm overrides and re-run extraction.
Review the current extracted products before validating selectors.
Use sampled product evidence to propose selectors with OpenAI.
Run validation first. Then confirm overrides and re-run Extract products.
Sample validation review
Review OpenAI’s suggested selectors before applying overrides.
Idle
Run validation to review sampled field suggestions.
The next step will appear here after OpenAI reviews the sampled product evidence.
No validation results yet.
Required fields
Run validation to inspect required field coverage.
Suggested selectors
Selectors ready to apply will appear here.
Suggested extra fields
No suggested extra fields yet.
Warnings & notes
Notes from validation will appear here.
Next step: Validate sampled fields (OpenAI)
Run validation to generate suggested selectors from sampled product evidence.
Run a step to see what it does and how it went.
Processing
Adjust snapshots and run extraction before moving to quality review.
Snapshot overrides
Quality workspace
All quality actions and review panels are grouped here: AI quality, Layer 1, and Layer 2.
Quality workspace
Run in order: AI quality check → Quality check (layer 1) → Second-layer quality (layer 2). AI stage updates now include whether jobs were reused or newly queued.
