MCP server · Speech to text

Talkthrough MCP server

Turn screen recordings, videos and audio into searchable transcripts, keyframes and OCR your AI can read.

Speech to text For it and development Multimedia Process

Talkthrough is a helper that lets your AI watch and listen to a recording for you. You give it a video or audio file on your computer, and it pulls out the spoken words, the on-screen text, and clear still frames. It is handy if you record bugs, meetings, demos or workshops and want notes without rewatching everything.

What is an MCP server? The 30-second version

On its own, your AI can only chat with you. An MCP server is a small helper program that gives your AI a new skill, like a connection to a specific tool or app. This one connects your AI to recordings on your own machine, so it can open a video, read the words and frames, and answer questions about what happened. Everything stays on your computer, and nothing gets posted anywhere.

What this MCP server does

You point your AI at a recording file, like a screen capture or a meeting audio. Your AI asks this helper to process that file. The helper listens to the audio, writes down what was said, grabs clear frames from the video, reads any text shown on screen, and lines everything up with real clock times. Then your AI can search it, quote it, and write you a summary or a draft report.

Flow diagram: you ask your AI “Turn my screen recording into a transcript and notes”, the Talkthrough MCP server works in steps: listen and write it down, then grab clear frames, then read the screen text, and you get back A transcript and key frames.

Click to zoom

What you can do with it

Turn a recording into a searchable transcript
Pull out key frames and read the text on screen
Find the exact moment someone said something
Label who said what in a meeting
Draft a bug report from a narrated screen recording
Summarize a workshop or demo into notes
Match remarks in a recording to system logs by time

Try asking your AI

“Process ~/Desktop/recording.mov and tell me what bugs were mentioned”
“Search this recording for the part about the login error”
“Give me the transcript of the meeting and list the action items”
“Pull the frames around the crash and read the error text on screen”

What it gives back to you

You get back a short summary of the recording, plus a transcript, key frames as image files, and the text found on screen. You can ask for just a slice of the transcript, a few frames near a moment, or a search result with the exact time. Your AI then writes the answer, notes or draft in the chat, quoting the recording.

Before you start

What you need

The uv tool installed (a small program that runs the helper; the README gives a one-line install)
A video or audio file on your computer, or a public video link
An MCP client like Claude Code, Claude Desktop, Cursor or Codex

Good to know

It reads the recording files you point it at, so only give it files you are comfortable having your AI look at.

Install it with your AI

Add Talkthrough MCP server to your AI, no technical skills needed

You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.

Sign in to get the install prompt

Members get a ready-made prompt that lets the Claude desktop app check Talkthrough MCP server, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.