A small setup.
Then, your own rhythm.

The development build has Reader, History and Settings. Integrated validation of these quick-read/history controls is still in progress. The recordings below show an earlier English interface: they do not show the new shortcut, language settings or history controls.

1. The toolbar and Ctrl+Shift+Y still open a local, silent preview. They never upload text or save a reading by themselves.
2. In Settings, connect your supported lab and choose the speech language. Interface language is separate: System, English, Português or Español. Changing it never translates the source or restarts narration.
3. Complete the first-use disclosure: choose to save readings on this device or read without saving. Setup or reconnection never replays an earlier Read request.
4. Select fresh text, then use Read selection aloud from the selection menu or its separately assigned shortcut (see its assignment in Settings). Check the actual assignment in Settings; if unassigned or conflicting, open Chrome’s shortcut settings.
5. Read authorizes the current selection in the saved speech language. Preparation takes time. If Chrome blocks playback, choose the prominent Play action. Repeated keys do not authorize duplicate requests.
6. To read different text, confirm replacement of the current reading. Owned work must finish or cancel first; an unknown submission outcome remains blocking.

chrome://extensions/shortcuts

History is a development-build feature whose integrated validation is still in progress. Optional local history starts only when you explicitly start reading, never when previewing. It stores exact selected text, a page title and filtered source address, speech language, confirmed position and measured playback. These may contain sensitive material. No credentials, audio or incognito readings are saved; nothing syncs to the cloud.

Use History to search, view an archive, open its safe source link, delete a reading, clear all or export versioned JSON. Export contains sensitive text: save it somewhere trusted. Turn off saving for future readings in Settings. Deletion cannot be undone.

Measured playback time is not proof of listening. Seeking is not listening; reaching the end does not prove full coverage. Checkpoints are saved about every two seconds and at playback transitions; a crash may lose the last few seconds. Reopening an archive never starts audio, resubmits a job or pretends to highlight the original page. Capture the source again and explicitly choose Play to continue.

History is bounded to 1,000 readings or 50 MiB, with up to 128 coalesced events per reading and visible compaction counts. Leitores never silently evicts readings. Capacity, migration and write failures are visible; a failed checkpoint pauses playback. Export or delete readings, or explicitly choose to continue without saving. Browser/profile removal or storage eviction can erase history: export is your backup.

1. On an ordinary webpage, select a visible passage in the main page. Start with one short paragraph.
2. Click the pinned Leitores icon to open Chrome’s native side panel. Review the capture, including the Exact selected source disclosure.
3. Choose the passage’s language: English, Portuguese or Spanish. This chooses how the text is read; it does not translate the selection.
4. When connected, press Play / Resume. This explicit action, with your language choice, authorizes the selected text to be sent to the lab.
5. Allow preparation to finish. If Chrome requires another playback gesture, press Play / Resume again.

Expected outcomePrepared audio plays and highlighting follows words on the original page. Preview alone is silent and local; speech preparation can take time.

Open capture ↗

Use ordinary top-frame HTML. PDFs, browser-internal pages, editable text, shadow DOM and frame selections are not supported. See exact selection limits ↓

Try Pause during prepared playback, then Play / Resume. Next, move the seek control. Notice that moving through the audio is separate from authorizing it to play.

Pause → Play / Resume

Pause retains the current playback clock. Play / Resume continues from that place rather than resetting the clip.

Seek

Choose a position within the prepared clip. Seeking leaves playback paused—even if audio was playing. Press Play / Resume when ready.

Previous / Next

Move between passages when available. Navigation stays paused pending an explicit Play / Resume; it is not automatic playback.

Stop

Stop resets the current clip to its beginning. It is not the same as releasing the captured selection.

Return to source

Restore the pinned source tab to see the passage in context. It does not capture a new selection.

Release selection

Let go of the pinned capture so another selection can be captured. Use this when you are ready for new text.

Expected outcomePause keeps the clock; seek and passage navigation wait for your next Play / Resume. Stop resets the current clip. Closing the side panel ends playback.

Open capture ↗

Open capture ↗

1. For another passage in the current capture, use Previous or Next when available, then explicitly press Play / Resume.
2. To find your current source page, choose Return to source. This restores the pinned source tab.
3. For different text, choose Release selection. Select a new passage on an ordinary page, then use the toolbar icon to capture it in the native panel.
4. Review Exact selected source again, choose the text’s language, and press Play / Resume only when you want it sent.

Expected outcomeThe new preview matches the new passage. Recapture itself does not authorize an upload. Navigation or a change to the source DOM invalidates the old selection; capture again rather than continuing with a stale source.

Open capture ↗

Open capture ↗

Start with the state you can see. These are written recovery steps, not fabricated error screenshots.