Gemini prompt

Gemini: Data Extraction from Screenshots

Extract structured data from screenshots, images, or visual documents into clean, usable formats.

Extract structured data from the attached image(s).

## What's in the image: {{image_description}}
## Desired output format: {{output_format}}
## Use case: {{use_case}}

### Extraction Instructions

**Step 1: Identify All Data Elements**
Look at the entire image carefully.
List every data element visible: tables, numbers, text fields, labels, headers, dates, etc.
Note any data that appears truncated, unclear, or partially visible.

**Step 2: Extract to Structured Format**
Convert all identified data to {{output_format}}.
For tables: preserve row/column structure.
For forms: extract field name + value pairs.
For charts/graphs: extract data points and labels.
For receipts/invoices: extract line items, totals, dates, parties.

**Step 3: Handle Ambiguity**
For any unclear text: provide your best interpretation + flag with [UNCLEAR: ...]
For missing data: use null or empty string + flag with [MISSING]
Do not guess at illegible numbers — mark as [ILLEGIBLE]

**Step 4: Quality Check**
Does the extracted data make logical sense?
Any totals that should add up — do they?
Any dates or numbers that look implausible?

**Step 5: Output**
Provide the extracted data in {{output_format}}.
Then provide: total fields extracted | unclear fields | missing fields | confidence level.

### Output Format Options
JSON: for programmatic use
CSV: for spreadsheet import
Markdown table: for readable display
Plain text key-value pairs: for simple forms

Fill in: {{image_description}}, {{output_format}}, {{use_case}}. In Loopy you fill these in a form and copy the finished prompt in one click.