Review & Fix

Overview · Enter Smart Review

After parsing you land in Smart Review, with four things keeping you in control:

Every change is recorded in history; open Sub-experiment → History to see the full change chain and undo in one click.

Edit the parse on the review page + review your changes

After parsing, the review page is now fully editable: experiment name, sub-experiments, every step (temperature / volume / reagents / code, etc.), every parameter; you can also add / delete sub-experiments, steps, and parameters. No need to create the experiment first and fix things afterward.

• Before [Confirm Create], required fields / duplicate parameter names are caught and highlighted on the spot
• If you changed anything, a "Review your changes" dialog lists exactly what you changed vs. the AI parse (old → new, added, removed) before it lands in the database

Unreviewed / Reviewed, view source, re-parse

• Every sub-experiment, procedure step and parameter has a marker on the right: an empty circle means Unreviewed, a green check means Reviewed. Clicking the row, editing it, or clicking the marker marks it reviewed; click the marker again to set it back.
• The top of the page shows how many items are left. Next unreviewed scrolls to the next one and moves focus there. Items that are partly in source, AI inferred, or missing from source come first; after those, the rest follow in page order.
• Your review marks are saved in this browser on this device. They survive a page refresh but not a different device or cleared site data. They only help you keep your place: they don't change any data and don't block confirming.
• View source: when you click a row that has a recorded source location (such as "Methods 2.1"), a source panel opens on the right on desktop, scrolled to that section with the match highlighted; on mobile a short excerpt opens under the row. The match is found by plain text search on the location and the row's own content, with no AI involved, and the page says so when nothing is found. Image imports keep no text copy of the source, so only the location is shown.
• Re-parse: if a parse went off track, click Re-parse next to the title. You can add a one-line hint such as "only the Western blot in Materials and Methods" (up to 200 characters, optional) and the same source is parsed again. It counts toward your usage like a new parse, and the current result stays as it is.

The review page for importing a single sub-experiment works the same way: the sub-experiment and each procedure step have provenance labels and review markers. "Sub-experiment name" is the name of this sub-experiment and "Procedure steps" are the steps under it; the small labels under a step say whether a value is a duration, temperature, volume or reagent.

8 finding types

The AI self-checks and flags each one, with these code identifiers:

• missing_param: should fill but didn't (concentration / temperature / duration)
• provenance_low: source unreliable (inference rather than verbatim)
• unit_mismatch: inconsistent units
• vague_step: too vague ("some / appropriate")
• unusual_value: outlier values (too high / too short)
• missing_control: no control / baseline
• contradicts_methods: contradicts Methods section
• spec_format: field format off (should be number not text)

4 provenance levels

• EXPLICIT: From the source (the paper states it directly)
• PARTIAL: Partly in source (the paper mentions it but lacks detail)
• INFERRED: AI inferred (filled in from context, with a citation)
• MISSING: Missing from source (not in the paper, you must fill it in)

On the review page, each step shows its source tag inline whenever it isn't "From the source" (Partly in source / AI inferred / Missing from source), so you can spot what needs checking at a glance.

Steps that came straight from the source are folded by default; expand them when you need to. Each label type gets a short one-time hint the first time it appears.