An idea pipeline, not a notes app

Speak it before the idea is gone.
Then let it grow.

Say it once — walking, driving, mid-conversation. It lands in your Obsidian vault as plain Markdown, where your agents can read it, connect it, and build on it. That is how a stray thought turns into something you ship.

iPhone · iOS 16+ · Android and desktop in the browser

Voice
Markdown
Your Vault
Your Agents
and the next idea starts sharper

Two Ways a Good Idea Dies

Walking

A brilliant idea pops up. Typing on your phone is too slow. By the time you get home, it's gone.

Commuting

Listening to a podcast sparks an insight. Your hands aren't free to write it down.

Three Weeks Later

You did write it down. It is sitting in some app with a few hundred others, and nothing ever pulls it back up.

Capture solves the first one. Only a pipeline solves the second.

Why This Actually Compounds

A pipeline is only as strong as its weakest link. Here is what holds at each of the three.

Capture happens on your phone

The tools you would compare us to run on a Mac, or run inside Obsidian — either way you have to be back at your desk with the app open. The idea does not wait for that. SpeakPen records where you actually are.

It lands as a file you own

Not a row in our database. A plain Markdown note in your own vault, through the SpeakPen Sync plugin in Obsidian's community store. Stop paying us and every note is still sitting there.

Your agents already read that folder

Claude Code, Codex, whatever you point at your vault — none of them need an integration with us. They need your spoken thought to exist as text they can read. That is the part we do.

From One Sentence to Something You Ship

1

Speak

  • Record on your phone — walking, driving, between meetings
  • Up to 20 minutes per recording (Pro)
  • Auto language detection, 99 languages
2

It Becomes a Note

  • AI transcription, cleaned up and punctuated
  • A summary and a title, written for you
  • Sorted automatically: meetings, lectures, notes
3

It Lands in Your Vault

  • Plain Markdown files, in your own Obsidian vault
  • SpeakPen Sync, from Obsidian's community plugin store
  • Or export the Markdown yourself — nothing is locked in
4

Your Agents Build On It

  • Point Claude Code or Codex at your vault — no integration needed
  • Or use the AI inside SpeakPen: ask, rewrite, pull action items
  • The thinking sharpens — and you have more to say

Useful Before Any Agent Touches It

Nothing to wire up. These work the moment the note lands.

Ask your notes

Chat across everything you've captured and get answers, with the source notes cited.

Instant action items

SpeakPen pulls the to-dos out of your note so nothing slips.

Rewrite in any style

Turn a rambly note into a polished email, blog post, tweet, or to-do list — in one tap.

Your Ideas, Flowing Everywhere

Notes arrive as plain Markdown files in your own vault. Stop paying and they are still yours — nothing stays trapped in our cloud.

Obsidian
Available
Claude · Codex · any agent
Via your vault
Copy / Share
Available
Notion
Coming Soon
Apple Notes
Coming Soon
Email
Coming Soon

Simple, Honest Pricing

Start free. Upgrade when you need more.

Free
$0
Experience voice capture
  • 3 recordings / month
  • Up to 5 minutes each
  • AI transcription & summary
  • Obsidian & Markdown sync
  • Copy & share notes
Get Started Free

Frequently Asked Questions

What is SpeakPen?
SpeakPen is a voice-first idea pipeline. You speak a thought into your phone, AI turns it into a clean, structured Markdown note, and that note syncs into your own Obsidian vault — where the AI agents you already use can read it and build on it. Capturing the thought is the first step, not the whole product.
How does the voice-to-text transcription work?
SpeakPen uses advanced AI speech recognition to transcribe your voice memos in real time. It supports 99 languages with auto-detection, converting your spoken words into clean, structured Markdown notes with summaries.
Can SpeakPen sync with Obsidian?
Yes. Install "SpeakPen Sync" from Obsidian's community plugin browser, then paste an API token — generate one in the SpeakPen web app at speakpen.app/app, under Settings → API Tokens. New notes then arrive in your vault as Markdown, each with a title, summary, and full transcript.
How is SpeakPen different from other voice note apps?
Most voice-to-Markdown tools for Obsidian either run on a Mac or run inside Obsidian itself, so the capture only happens once you are back at your desk with the app open. SpeakPen captures on your phone, at the moment the idea shows up, and the note lands in your vault afterwards. That is the difference: it is there when your Mac is not.
What languages does SpeakPen support?
Automatic language detection covers 99 languages — just speak and SpeakPen figures it out. If you prefer to pin a language manually instead of relying on detection, the manual picker currently offers 10: English, Chinese (Simplified and Traditional), Japanese, Korean, Spanish, French, German, Portuguese, and Italian.
Is SpeakPen free to use?
SpeakPen offers a free plan with 3 recordings (up to 5 minutes each) so you can experience the full pipeline. The Pro plan at $9.90/month unlocks 60 recordings per month with up to 20 minutes each.
Can I import existing audio files?
Yes, SpeakPen supports importing audio files in addition to recording directly. Imported audio goes through the same AI transcription and summary pipeline as live recordings.
Do I need Claude or Codex for this to be useful?
No. Transcription, summaries, action items, rewriting, and asking questions across your notes all happen inside SpeakPen. The agent step exists because your notes are plain Markdown in a folder you control — anything you already point at that folder can read them, so nothing has to be integrated on our side.

Start the Loop With One Recording

Free to start. Say one thing into your phone and see where it ends up.

Download Free on iOS