Turn a Screenshot into Tasks
Turn a screenshot, a photo of your notes or a whiteboard into tasks. GetItDone's AI proposes tasks with subtasks and dates, and you approve each one.
Capture turns a picture of your work into tasks. Drop in a screenshot of a Slack thread, a photo of your notebook, or a whiteboard from the last meeting, and GetItDone proposes the tasks it finds: names, descriptions, priorities, dates, and subtasks. You review every card before anything lands in your task list.
Start a capture
Choose Capture tasks. It is on the Daily board header next to Quick add and Plan my day, in the tasks toolbar, and on the empty "no tasks yet" screen.
Then give it an image:
- Drop a file onto the capture area.
- Paste a screenshot. Click the capture area first; paste works while it has focus.
- Choose a file, or use Take a photo on a phone to open the camera.
One image per capture. Large images are shrunk in your browser before upload, so a full-resolution phone photo is fine.
Review the proposals
While it reads, you see "Reading your capture…". Each task it finds becomes a card, and every card is editable:
- Task name and description.
- Priority: none, Low, Medium, High, or Urgent. The AI sets a priority only when the image says how urgent something is.
- Due and Start dates. The AI never invents a date that is not in the image.
- Subtasks, one level deep. Rename them as you like; a subtask whose name you clear is skipped.
Every card starts checked. Untick the ones you do not want, then choose Add N tasks. Start over and Dismiss leave your task list untouched.
A capture proposes at most 10 tasks, each with at most 10 subtasks. Duplicates and unnamed proposals are dropped before you see them.
What gets created
For each checked card, GetItDone creates the task, sets its description, and creates its subtasks under it. The original image is attached to every task the capture created, so the source stays one click away.
Captured tasks go to your task list. They are not placed on today's Daily board automatically; add them to your day like any other task.
Capture only ever proposes. Text inside the image cannot create, change, or delete anything by itself. The model's output is checked and capped before you see it, and you approve every task.
When nothing is found, or the AI is unavailable
If the image has no tasks in it, you see "We couldn't find any tasks in that capture". If the AI cannot read it, the sheet offers Type your tasks instead: one task per line, then Add N tasks. After an error, Retry sends the same image again, so you do not have to find it twice.
AI actions and plan limits
Each capture sent to the AI uses one AI action from your plan's allowance, including one whose read then fails; a retry is a new capture. See Pricing for the allowance on each plan. When the allowance runs out, capture shows an upgrade prompt; typing tasks in yourself never costs anything.
The image upload counts against your plan's per-file limit like any attachment (25 MB on Free).
Where your image goes
The image is uploaded to the same storage as your task attachments. Once you approve a capture, it is an ordinary attachment and lives as long as the task does. Images from captures you never approve are deleted after 30 days.
Voice notes
Voice capture (record up to two minutes and turn it into tasks) is built but not enabled yet. Until it is, the capture sheet points you to a screenshot or typed tasks instead.
From the browser extension
The browser extension can attach a screenshot of the page you are on to a single new task. It does not run AI extraction; to turn a screenshot into several tasks, use Capture in the web app.
Related
- Quick add in plain language: type a task as a sentence, with dates, priority, project, and repeat rules.
- Use GetItDone in Claude: let an AI assistant read tasks, including their attached images.
Read this page as Markdown — plain text for AI agents. Every guide at once: llms-full.txt.