Speak
Tap once and say it while it is still clear. Your device's own speech engine writes it down — no upload, no waiting on a server.
ANDROID BETA LIVE · iOS IN DEVELOPMENT
Say it once.
It lands where it belongs.
Talk to Action is the front door to your second brain. Tap once, say the thought, and it lands in Obsidian, Google Keep or your clipboard — without opening a single one of them.
A fast voice capture inbox for the tools you already use — available on Android today.
Free · No account · No sign-up · 16 MB APK · Google Play listing lands after developer verification
Latest draft
Destination
Obsidian
Intent
Ready
Note
Task
Shop
Tap and speak
- 1 tap from home screen to recording
- 0 accounts — nothing to sign up for
- On‑device transcription and sorting
- Free tier is a complete tool, not a trial
01 / The moment of capture
Good systems fail
at the first step.
By the time you have unlocked the phone, found the right app, waited for it to load and decided which folder this belongs in, the thought is gone. Every note-taking system on earth is downstream of one question: how fast can you get it in?
02 / The flow
Speak it. Shape it.
Send it.
Shape
You get an editable draft, already tagged Note, Task or Shop by an on-device classifier. Wrong guess? One tap fixes it. Capture never waits for sorting.
Send
Fire it at the destination that keeps the work moving — your vault, your list, your clipboard, or an AI you want to think out loud with.
03 / Your capture layer
The fastest way
into Obsidian.
Talk to Action doesn't want to replace your vault — it wants to feed it. Captures are appended straight into your daily note through Obsidian's own URI handler, with your date and time format, in the folder you already use. Not a share-sheet dump. A real append.
Different route per category, too: Notes to your daily note, Shop to your shopping list, Tasks wherever tasks live.
Obsidian is a trademark of Dynalist Inc. Talk to Action is an independent project.04 / The system behind the capture
Your AI second brain
is only as good as what reaches it.
Talk to Action is not another place to keep information. It is the fast, quiet entrance that feeds the system you already use to remember, retrieve, and think.
Personal knowledge workflow
↗
Obsidian + Claude
Catch the raw thought before it fades. It enters the vault you trust, where it can become durable context for the questions you explore with Claude.
- Speak a thought
- Append it to your vault
- Explore the growing context
Living research workflow
↗
Google Docs + NotebookLM
Keep a running document of ideas, meetings, and observations. Add it to NotebookLM or Gemini Apps when you want grounded research, synthesis, and a conversation with your own sources.
- Speak it into a living document
- Build a source you own
- Ask better questions of it
Brand marks identify products that can sit alongside a Talk to Action workflow. Talk to Action is independent and is not affiliated with or endorsed by their owners.
05 / One inbox, many exits
One voice inbox.
Your stack beyond it.
Your thought should land wherever the next step belongs. Choose the destinations you use, then send without rebuilding the thought in every app.
Knowledge & notes
- Obsidian direct append into your vault
- Google Keep note or checklist
- Google Docs build a living document
- Evernote
- Clipboard take it anywhere
Think & get things done
- ChatGPT · Claude · Gemini start an AI conversation with the thought intact
- Todoist · TickTick
- Motion
- Google Calendar · Google Tasks
- Microsoft To Do
Teams & automation
- Slack
- Microsoft Teams
- Webhook connect Zapier, Make, or your own endpoint
17 destinations are ready today. A destination only receives text when you choose to send it there — nothing syncs in the background.
A pause, before the system
A thought is small.
What it becomes is not.
Capture is only the first movement. The next is choosing the system that can turn one clear thought into context, action, or a better question.
06 / Platforms
Android now.
iPhone is being built.
Both apps are written natively — Kotlin on Android, SwiftUI on iOS — because the speech engine and on-device processing live deep in each OS. No cross-platform shell.
Android
Beta available now
- Install the signed beta APK directly, today
- Google Play listing follows developer verification
- Looking for closed-test participants right now
iOS
In active development
Running on device
The SwiftUI build launches on a physical iPhone. Capture, review, destinations, history and settings are implemented.
Feature parity with Android
Matching the Android build screen for screen, including the share sheet and the Obsidian URL hand-off.
TestFlight beta
Opens once parity testing passes on real hardware.
App Store
Review materials and privacy disclosures follow the beta.
- Apple Notes and Reminders arrive with the iOS build
- Apple Watch capture is planned as a later Pro extension
- The list is the only way to hear the moment TestFlight opens
07 / The honest comparison
You already have
a microphone button.
So the real question is what a dedicated capture app earns you. Here is the case, stated plainly.
| What you want | Keyboard dictation | Assistant Siri / Google |
Talk to Action |
|---|---|---|---|
| Start without choosing an app first | No — open the app, then dictate | Partly | Yes — one tap, then talk |
| Edit before it is committed | Yes | Rarely | Yes — always an editable draft |
| Land inside Obsidian | Manual copy-paste | No | Direct append |
| Sorted into note / task / shopping | No | Fixed to its own apps | On-device, one-tap correction |
| Several destinations at once | No | No | Yes (Pro) |
| Works with no account | Varies | Needs the platform account | Yes |
08 / Privacy
Your thought doesn't
go on a trip.
Recording, transcription and classification all happen on your phone. There is no Talk to Action server, no account, and no analytics profile. Text leaves the device only when you send it, and only to the destination you picked.
- Microphone Used while you are recording. That's the whole story.
- Speech recognition Your device's built-in engine, in whatever language your voice input is set to.
- Network Only touched when a destination you chose is a cloud service.
- Pro AI features Bring your own API key. You pay the AI provider directly; nothing routes through us.
09 / Pricing
Free is a tool,
not a trial.
The complete loop — speak, transcribe, correct, send — is free forever, with no time limit and no account. Pro is for people who route to more than one place.
Free
$0
Forever. No sign-up.
- Unlimited voice capture
- On-device transcription
- Note / Task / Shop sorting with one-tap correction
- Editable draft before every send
- One active destination
Pro
Not yet on sale
Arrives after the Play release.
- Multiple destinations, sent simultaneously
- Per-category routing rules and templates
- Full history: search, restore, resend
- Custom categories and work / personal profiles
- Home-screen widgets and shortcuts
- BYOK AI: high-accuracy speech, task extraction, cleanup and translation
10 / Questions
The things
you'd ask first.
Is installing an APK safe?
It's the same build that goes to Google Play, signed with the same release key — Play verification is an account formality, not a code review we failed. Android will still ask you to allow installs from your browser, which is the system working as intended. If you'd rather wait for the Play listing, join the tester list and you'll get the link the day it opens.
When is the iPhone version out?
It already builds and runs on a physical iPhone. What's left is matching the Android app screen for screen, then real-device testing, then TestFlight. No date is promised until parity testing passes — join the iOS list and you'll hear the moment it does.
Does it work offline?
Capture, transcription and sorting do, as long as your device's speech engine has its offline language pack. Sending to a local destination like Obsidian or the clipboard needs no connection either. Cloud destinations obviously need one.
Which languages does it understand?
Whatever your phone's voice input already handles — the app uses the OS speech engine rather than shipping its own, so it inherits that list. English and Japanese are what's been tested end to end so far; more are being verified before they're claimed.
How is this different from a voice memo app?
Talk to Action is built for the next step after capture: review the transcription, choose or correct its type, then send it to the note, task, AI, or automation tool you use. A voice memo is simply recorded; this is prepared for action.
Do I need an account?
No. There's no sign-up, no profile and no server to hold one. Settings live on your device.
Get started
Your next thought
deserves a way in.
Install the Android beta now, or get one email when the iPhone build reaches TestFlight.
Wanted: Android closed testers
Google Play needs 12 testers running the app for 14 days before it can go public. If you capture into Obsidian and want the app to exist, this is the single most useful thing you can do — and you get every build first.
Put me on the tester list