Skip to main content

TellDone on the web

The web app at app.telldone.app is a full companion to the iPhone app. Same account, same data, same voice-first pipeline. It works in any modern browser (Chrome 90+, Safari 14.3+, Firefox 90+, Edge 90+) and can be installed as an app on iOS, Android, and desktop.

First run, in 60 seconds​

  1. Open app.telldone.app.
  2. Sign in with email + password, Sign in with Apple, or Sign in with Google.
  3. After first sign-in, a short Welcome carousel plays. Tap Start recording at the end.
  4. Your first tap on the large mic button shows a brief "Allow microphone access" hint. Tap Allow mic.
  5. The browser asks for microphone permission - choose Allow (not Allow once).
  6. Say something simple: "tomorrow at 3pm doctor's appointment, and remember to call mom". Tap the mic button again to stop.
  7. Within 5-30 seconds, your note appears with an extracted task, event, and a short summary.

When installed as a Progressive Web App, TellDone lives on your home screen without the browser URL bar and keeps microphone permission between sessions. On iOS this is the only way Safari will remember the mic permission - otherwise it re-asks every time.

iOS (Safari)​

  1. Open app.telldone.app in Safari (Chrome on iOS does not support PWAs).
  2. Tap the Share icon at the bottom.
  3. Scroll down and tap Add to Home Screen, then Add.
  4. Launch from the home-screen icon - it opens in full-screen mode.

Android (Chrome)​

  1. Open app.telldone.app in Chrome.
  2. An Install TellDone banner appears after a few seconds. Tap Install.
  3. If the banner doesn't show, tap the menu (three dots) > Add to Home Screen.

Desktop (Chrome or Edge)​

  1. Open app.telldone.app.
  2. Look for the Install icon (a small box with an arrow) on the right side of the address bar.
  3. Click it and confirm Install. The app opens in its own window without tabs.

Recording in the web​

Single tap​

Tap the red mic button, speak, tap again to stop and save. The default mode on any device.

Push-to-talk (press and hold)​

Press and hold the mic button, speak while holding, release to save. Good for short thoughts.

Drag to cancel​

Press the mic button and, while holding, drag your finger far to the side - the recording is discarded without upload. Useful when you started by accident.

Keyboard shortcuts (desktop)​

  • R - start or stop recording (when no input is focused)
  • Space - stop while recording
  • Esc - cancel the current recording

Type instead of speaking​

Tap the Type instead button next to the mic. You get a keyboard with Paste, Cancel, and Send. The text goes through the same AI analysis as a voice recording.

What happens after you send​

A small chip above the mic shows the pipeline stage:

ChipWhat is happening
Uploading...Audio is sent to the server
Transcribing...Speech becomes text (Soniox). ~5-30 seconds per 10 minutes of audio
Understanding context...AI extracts title, summary, tasks, and events
Finding tasks...Saving to the database, matching against existing tasks (Smart Context)

When the chip disappears and a toast pops up with the note title, processing is done. Long recordings (over 2 minutes) get a small ✨ next to the chip - they take a bit longer. If a stage stays more than 45 seconds, the chip adds "still working" - normal under peak load. A one-hour recording can take up to ~5 minutes end to end.

Follow-up recordings​

Link a new recording to an existing note, task, or event so the AI has full context and can update it directly.

  1. Open any note, task, or event (tap its card in the list).
  2. In the right-side panel (or bottom on mobile), tap Follow-up.
  3. A chip appears at the top: "In response to: <title>". Record or type as usual.
  4. The new note is saved with a link to the parent and appears in the parent's detail view.

To clear the parent link before recording, tap the ✕ on the chip.

See Follow-up recordings for the full list of actions the AI can take on each parent type.

Main pages​

PageWhat's there
NotesYour voice and text notes. Detail panel: audio player, transcript, extracted tasks and events, tags, follow-up
TasksFiltered list (todo / done / archived), one-tap complete toggle
EventsCalendar plus list view, reminders
ReportsDaily, Weekly, Monthly, Yearly with period switcher
TagsPin, remove, rename; AI-generated tags mixed with manual ones
ActivityUpload history with stages and statuses - useful if something looks stuck
SettingsProfile, account, integrations, MCP, themes, and more

Settings overview​

All settings are available in the web app:

  • Profile - display name, email verification status
  • Account - change email, change password, Apple/Google sign-in status, promo-code redemption
  • Email reports - daily, weekly, monthly, yearly toggles; note-forwarding address with email verification
  • Appearance - light, dark, or system theme; accent colour (Ultra-exclusive themes on the Ultra plan)
  • Voice and language - UI locale, transcription locale (can be different), timezone
  • Smart AI - profile synthesis, Smart Context autocomplete, smart achievements
  • Notifications - reminder push toggle, forward-email verify
  • AI Agents (MCP) - enable for Claude Code, Cursor, Windsurf, Codex, OpenClaw; copy or regenerate the token; per-client setup snippets
  • Integrations - Todoist, Notion
  • Data and privacy - data export, delete account, replay onboarding
  • About - version, help / privacy / terms links, contact support

Troubleshooting​

"Processing..." chip hangs for a long time​

Possible causes:

  • A long recording (up to ~5 minutes for a one-hour file, which is normal)
  • A short load spike on Soniox or OpenAI (usually 1-2 minutes, clears on its own)
  • A dropped WebSocket connection (laptop went to sleep, VPN switched)

What to try:

  • Wait 3 minutes - the web client checks the note status on the server and clears the chip.
  • If it still hangs, reload the page. The client re-syncs from the server.
  • If the note still does not appear, check the Activity page. You will see the stage (uploaded / transcribing / analyzing / completed / failed) and any error.

The browser asks for microphone permission every time​

On iOS Safari, this happens in a regular tab - it is the browser's behaviour. Install the PWA (see Install as an app above) and the permission sticks.

On other browsers, choose Allow (not Allow once or Allow this time). If you made the wrong choice, reset site settings:

  • Chrome / Edge: lock icon in the address bar > Site settings > Microphone > Reset
  • Firefox: lock icon > Microphone > Clear permission
  • Safari (Mac): Safari > Settings > Websites > Microphone > app.telldone.app > Allow

"Microphone blocked"​

Means the browser previously got a Block answer. Only fixable in site settings (see above) - the web app cannot unblock itself.

"Plan limit reached"​

You used up the monthly limit (recordings, STT minutes, or notes). Limits reset at midnight in your timezone. Upgrade in Settings > Account, or wait for the reset.

"Silent recording" or "Too short"​

The recording was empty or under ~1 second. Speak closer to the microphone, check that your system mic is not muted, and try again.

"Couldn't save recording for retry"​

Browser storage is full or you are in private mode. Leave private mode or clear some browser cache and retry.

Need to delete data​

  • One item - use the delete button in the detail view. Soft-deleted items can be restored in Settings > Data and privacy > Trash during the grace period (7 days on Free, 30 on Basic, 90 on Pro, 365 on Ultra).
  • Full account - Settings > Data and privacy > Delete account. A grace period applies; you can cancel within it.

What the web app doesn't do yet​

Some iPhone app features are not in the web app yet:

  • In-app subscription management - use the App Store, or contact support
  • Full webhook automation editor - you can enable automations but some fine-tuning is iOS-only
  • Web push notifications - only local reminder scheduling
  • Archived task filter - only active and done

We're working on these.

See also​