The screenshot-to-article feature lets you upload images of a workflow and have the AI write the full article for you — headings, numbered steps, and each screenshot placed inline where it belongs. It's the fastest way to document a feature you can show but don't want to write from scratch.
This is a PRO feature. The button is visible on all plans but locked unless your workspace is on PRO.
How to use it
Step 1 — Open the modal
Open any article in the editor (new or existing). In the AI Assist bar between the metadata fields and SEO Settings, click Generate content from screenshots.
Step 2 — Upload screenshots
Drag and drop your images onto the upload area, or click it to browse. Images are processed in the order you add them — Screenshot 1, Screenshot 2, and so on — so add them in the sequence a reader would follow.
Supported formats: PNG, JPG, WEBP
Per-file limit: 2 MB
Maximum images: 5 per generation
Images are automatically compressed (resized to max 1280 px on the longest edge) before being sent to the AI.
Step 3 — Add context (optional but recommended)
The Topic / context field lets you tell the AI what the article is about. Without it, the AI infers the topic from the screenshots alone — which usually works, but a short hint produces more accurate titles and step descriptions.
Good context examples:
How to invite a teammate to your workspaceSetting up two-factor authenticationExporting a report to CSV
Keep it under a sentence. The field accepts up to 1,000 characters.
Step 4 — Generate
Click Generate article. The AI reads every screenshot and writes the full article — title, description, and markdown content with each image embedded at the relevant step.
This usually takes 10–30 seconds depending on the number of images. You can cancel at any time during generation.
Step 5 — Preview and insert
Once generation finishes you'll see a preview of the title, description, and formatted content. Review it before committing.
- Insert into editor — writes the generated content into the article editor. If the article already has content, you'll be asked to confirm before it's replaced.
- Regenerate — goes back to the upload screen so you can adjust the context or swap images and try again.
After inserting, the article is still a draft. Review, edit, and publish when you're ready.
Limitations
| Constraint | Detail |
|---|---|
| Max screenshots | 5 per generation |
| File size | 2 MB per image |
| Formats | PNG, JPG, WEBP only |
| Generation time | 10–30 seconds |
| Title length | Capped at 70 characters |
| Description length | Capped at 160 characters |
| Existing content | Inserting replaces the current editor content entirely |
The AI generates output from what it can see. If a screenshot is cropped tightly or shows only a partial UI state, the written step may lack context — you'll need to fill that in manually.
Best practices
Order your screenshots before uploading. The AI numbers them 1, 2, 3… and places them in the article in that order. Getting the sequence right before uploading saves you editing time.
Crop to what matters. A screenshot showing only the relevant dialog or panel is far more useful to the AI than a full-screen capture with unrelated UI around the edges.
Use the context field. Even one sentence — the article title you have in mind — meaningfully improves the output. The AI uses it to write step descriptions that match your intended framing.
Treat the output as a first draft. The AI writes clear, structured prose but it doesn't know your product's terminology, tone of voice, or edge cases. Budget a few minutes to edit step descriptions, adjust the title, and add any caveats the screenshots don't show.
Start with a blank article. Generating into an article that already has content replaces everything. If you want to add screenshots-based content to an existing article, duplicate it first or copy the existing content somewhere safe before generating.
Last updated: June 6, 2026