GitHub - EdonZo/talkthru · GitHub
/" data-turbo-transient="true" />
Skip to content
Type / to search
Sign in<br>Sign upAppearance settings
You signed in with another tab or window. Reload to refresh your session.<br>You signed out in another tab or window. Reload to refresh your session.<br>You switched accounts on another tab or window. Reload to refresh your session.
Dismiss alert
{{ message }}
EdonZo
talkthru
Public
Notifications<br>You must be signed in to change notification settings
Fork
Star
main
BranchesTags
Go to file
CodeOpen more actions menu
Folders and files<br>NameNameLast commit message<br>Last commit date<br>Latest commit
History<br>22 Commits<br>22 Commits
.github/workflows
.github/workflows
fixtures
fixtures
packages
packages
scripts
scripts
site
site
.gitignore
.gitignore
LICENSE
LICENSE
README.md
README.md
package-lock.json
package-lock.json
package.json
package.json
View all files
Repository files navigation
talkthru
talkthru.dev · npm · MIT
An automated feedback loop for any app. Screen-record your app on your phone and talk<br>while you use it. Send it to your machine and your coding agent gets everything you<br>said, attached to the screen you were looking at when you said it.
you: *tapping through checkout* "this button is too small, I keep missing it"<br>"and this error doesn't say what went wrong"
agent: sees both screens, finds both components, fixes them.
Runs entirely on your machine. No accounts, no uploads, no API keys.
Works with a mac screen recording (⌘⇧5) or an iPhone one over AirDrop. Any video with sound works — point the watcher at a folder.
70-second demo — install, record, and the agent picking it up.
Install
npx talkthru doctor --fix
Pulls ffmpeg, whisper.cpp and the speech model. Nothing else to set up.
Start the watcher
npx talkthru watch
Leave it running. It only touches files named like a screen recording — the rest of your<br>Downloads folder is invisible to it.
Connect your agent
talkthru is an MCP server, so any MCP client works — Cursor, Windsurf, Cline, Zed,<br>Claude Desktop. Add it to your client's config:
"mcpServers": {<br>"talkthru": {<br>"command": "npx",<br>"args": ["talkthru-mcp"]
Claude Code has a one-liner for the same thing:
claude mcp add --scope user talkthru -- npx talkthru-mcp
Restart your client afterwards so it picks up the server.
Built and tested against Claude Code. Other MCP clients should work — tell me if yours doesn't.
Set up your mac
Once. Press ⌘⇧5 , pick a record mode (the icons with the ◉ dot), then open<br>Options and set Save to → Downloads and your Microphone . Both stick.
1 · 2. press ⌘⇧5 , pick a record mode, open Options
3. Save to → Downloads — the folder talkthru watch looks at<br>4. pick your mic under Microphone
5. hit Record and talk as you use your app
No mic selected means a silent video and nothing to transcribe. Worth checking before<br>your first recording.
Set up your phone
Optional — only if you want to record a phone app rather than your desktop.
Once. Control Centre → + → Add a Control → Screen Recording . Then press and<br>hold it and turn the microphone on .
1. tap +<br>2. Add a Control<br>3. pick Screen Recording
4. press and hold ◉<br>5. Microphone On<br>6. tap ◉ and talk
Use it
Record your app, talk as you go
Get the video onto your machine (AirDrop is easiest)
Ask your agent: "check the feedback in my last talkthru session and implement it"
Ready in about twenty seconds. Your original video is kept in ~/.talkthru/archive/.
Mute your app if it talks. Its narration lands in the transcript next to yours and<br>whisper can't tell you apart.
Output
Your words, attached to the screen that was up when you said them:
## 00:34 · f11 — `frames/f11.jpg`<br>- [00:39] "this button is too small, I keep missing it"
## 00:52 · f13 — `frames/f13.jpg`<br>- [00:55] "and this error doesn't say what actually went wrong"
Plus the frames. A two-minute session is about 600 tokens.
Commands
talkthru watch # watch ~/Downloads for screen recordings<br>talkthru watch ~/some-folder # watch your own folder, any video<br>talkthru process video.mp4 # process one file<br>talkthru list # recent sessions<br>talkthru show latest # print one<br>talkthru doctor # check the setup<br>talkthru prune # delete old sessions
tt also works.
How it works
ffmpeg pulls the frames where your screen actually changed. whisper.cpp transcribes<br>your voice locally, split on the real silence between sentences, so each thing you said<br>lands on the screen you were looking at. An MCP server hands it to your agent.
Needs Node 20.11+.
Why
I built this because I test my own app on my phone every day and kept losing the feedback<br>on the way back to my editor — screenshotting, cropping, trying to describe which screen I<br>meant. Now I just record, talk, and Claude Code picks it up. Sharing it in case it saves<br>you the same trip.
About<br>No description, website, or topics provided.<br>Resources<br>Readme<br>MIT license<br>Activity<br>Stars<br>2 stars<br>Watchers<br>0 watching<br>Forks<br>1 fork<br>Report...