Where does zoom_read_transcript look for the transcript?

It reads the text artifacts saved alongside your local Zoom recordings — the .vtt caption track or closed_caption.txt file in your Zoom recordings folder. The recording must already be saved on your Mac with captions or a transcript enabled.

Does it upload my Zoom recording anywhere?

No. The transcript file is read and cleaned entirely on your Mac. Nothing is sent to an LMCP server or any cloud service, and nothing is stored remotely.

What if my recording has no transcript file?

The tool reads existing text artifacts (.vtt or closed_caption.txt). If Zoom didn't generate captions or a transcript for that recording, there's no text to read — enable audio transcript or closed captions in Zoom before recording.

Which AI assistants can use this tool?

Any MCP client: Claude Desktop, Cursor, VS Code, Windsurf and Zed auto-configure via local stdio, while ChatGPT, Claude.ai, Grok and Perplexity connect through the LMCP Cloud Relay connector.

Read Zoom Transcripts with AI on Mac

zoom_read_transcript reads the text artifacts of a Zoom recording already saved on your Mac. It opens the recording's transcript or captions file (a .vtt caption track or closed_caption.txt) and returns it as clean, readable text — timestamps and speaker noise stripped out so your AI can actually work with the words. Nothing is uploaded; the tool reads the files sitting in your local Zoom folder.

A typical prompt you might type: "Read the transcript of my last Zoom call and pull out every action item assigned to me." Your AI calls zoom_read_transcript, gets the full conversation as text, and hands you a tidy summary in seconds.

Which AI agents work?

LMCP exposes 150+ tools to any MCP-capable assistant. Desktop clients auto-configure over local stdio — no manual setup:

Claude Desktop
Cursor
VS Code (GitHub Copilot)
Windsurf
Zed

Web-based AIs connect through the LMCP Cloud Relay connector, which bridges them safely to your Mac:

ChatGPT
Claude.ai (web)
Grok
Perplexity

Automation

zoom_read_transcript shines when chained with other LMCP tools. Read a transcript, then ask your AI to draft a follow-up email in Mail to the attendees, create Reminders for each action item, add a Calendar event for the next sync, or drop a summary into Notes. One prompt like "Summarize today's Zoom standup, email the notes to the team, and set reminders for my tasks" can fan out across Mail, Reminders, Calendar and Notes — all from the transcript text the tool extracted.

Context

Because LMCP runs on your machine, the AI works with your real recordings — the meetings you actually had, in the folder where Zoom saved them. There's no copy-pasting walls of caption text, no exporting, no re-uploading to a web tool. The assistant reads the genuine artifact and reasons over the exact words spoken, with the names and details that matter to you.

Productivity

Manually scrubbing a 45-minute Zoom transcript for decisions and to-dos is tedious and error-prone. With zoom_read_transcript the AI reads the whole thing instantly and returns a clean digest — key decisions, owners, deadlines — in the time it takes to ask. Turn a recorded call into shareable meeting notes, a status update, or a task list without opening the recording yourself. Download LMCP and let your AI handle the read-through.

Privacy & GDPR

Everything happens locally. The transcript file is read on your Mac, processed on your Mac, and never copied to any LMCP server or third-party cloud. There are no API keys to hand over and nothing is stored remotely. This is privacy GDPR-compliant by architecture — your meeting content stays yours.