For Linux developers · Free desktop preview

Keep the error.
Find it by its text.

Capture terminal output, code snippets, and debugging references. CaptureRecall searches their visible text on your device, so you can return to the screenshot without remembering its filename.

Download for Linux ↓Try the workflow →

v0.3.5 · Ubuntu / Debian x86_64 recommended · No account required

A library for the captures you choose

From a disappearing error to a reusable reference.

A terminal error, failing build, or configuration screen can be useful long after it disappears. Capture the relevant area, let local OCR recognize its text, and add a project note if you need more context.

Capture	Try searching	Add context
Terminal error	`permission denied`	The command and project
SQL example	`CREATE INDEX`	Why you saved the query
Configuration screen	A visible setting label	The environment and intended change

These examples illustrate the workflow. Search matches recognized text and saved metadata; recognition depends on image quality.

Make your first useful capture

Capture. Search. Reuse.

Install the matching package. Download the x86_64 DEB for Ubuntu or Debian, verify its checksum, and install from your Downloads folder:
```
sudo apt install ./CaptureRecall_0.3.5_amd64.deb
```
RPM and AppImage alternatives are on the download page.
Use harmless sample text. Show permission denied in a terminal or document. Capture that area in CaptureRecall and approve your desktop's capture prompt. Keep passwords, tokens, and private messages out of this trial.
Let local OCR finish. Create two more captures with different visible text. This gives you a small library to search rather than a single image.
Find the first capture. Enter permission denied in library search. Open the result and check it against the original PNG. If OCR misses the phrase, use larger, clearer text or add it as a note.
Bring it back into your work. Copy the image or drag its saved PNG into an application that accepts images. Add a project tag or note so you can find it again.

What counts as a useful first trial?

You saved three captures, found the right one by a visible phrase, and copied or dragged it into another app. If any step fails, that is useful feedback too.

Check before installing

A preview with clear boundaries.

Linux packages target x86_64. Ubuntu and Debian are the recommended starting point; other distributions are best effort.
Capture uses the XDG Desktop Portal. Desktop environment, Wayland/X11 session, portal configuration, and permissions can affect capture behavior.
OCR currently targets Latin-alphabet text and can miss small, blurry, or low-contrast words.
Search looks for text in your CaptureRecall library. There is no bulk import of your existing screenshot folders, automatic screen recording, or natural-language visual search.
Local OCR does not require optional vision analysis. That separate feature is off by default and needs a trusted local service.

Read the privacy model, OCR guide, and published release notes. Apple Silicon Mac preview packages are also available, but the DMG is not Apple-notarized.

Help shape the next release

Where did the workflow help or break?

If you want to share feedback, tell us your distribution, desktop environment, package type, and which step succeeded or failed. Would you use this for debugging references? Would importing existing screenshots make it more useful?

GitHub issues are public. Do not attach private captures, secrets, or identifying log output. The app does not send trial reports or analytics automatically.

Share feedback on GitHub ↗