Claude MCP
Meeting
Transcription
Connect Claude Desktop, ChatGPT, or any MCP-compatible agent to live meeting transcription. Your agent can drop a bot into a meeting, stream transcripts, and act on what's being said — all via MCP tools.
Waiting for meeting participants…
Feed meetings into your AI. Live.
AI needs data, and the best data is real-time. Vexa streams transcripts, audio, and context from any meeting straight to your models and agents — as it happens.
Built for MCP-native agents
Vexa is the only Apache-2.0 meeting transcription API that ships as an MCP server. Add one entry to your Claude Desktop config (or any MCP-compatible host) and your agent gets a `request_meeting_bot` tool, a `get_meeting_transcript` tool, and live access to what's happening in any Google Meet, Microsoft Teams, or Zoom call. Build agents that join meetings on your behalf, summarize as they go, draft follow-ups in real time, and trigger CRM / Slack / Linear actions when the transcript matches a rule.
- Remote MCP server — one command in Claude Code, or one connector URL in Claude Desktop
- Nine tools including request_meeting_bot, get_meeting_transcript, list_meetings — agents read live meeting context
- Same Apache 2.0 license: run the MCP server alongside your self-hosted Vexa for full data residency
claude mcp add --transport http vexa https://api.cloud.vexa.ai/mcp \
--header "Authorization: Bearer $VEXA_API_KEY"
# Claude Desktop / claude.ai: Settings -> Connectors -> Add custom connector
# URL: https://api.cloud.vexa.ai/mcp
# Then in Claude:
# "Drop a bot into https://meet.google.com/abc-defg-hij,
# summarize the conversation every 5 minutes, and flag
# any mention of pricing or renewal."Set $API_BASE to https://api.cloud.vexa.ai (hosted) or http://localhost:8056 (self-host). Get $API_KEY from /get-started.
Claude MCP integration FAQ
What agent builders ask before they wire Vexa into Claude / ChatGPT / custom MCP hosts.
What MCP hosts does this work with?
Any MCP-compatible client — Claude Desktop, Claude.ai (via remote MCP), ChatGPT (via its MCP-compatible flow), Continue.dev, and any custom agent built with the official MCP SDK. Vexa is a standard REMOTE MCP server over streamable HTTP, so any host that supports remote MCP can connect to it directly.
What tools does the MCP server expose?
Nine: parse_meeting_link, request_meeting_bot, get_bot_status, update_bot_config, stop_bot, list_meetings, get_meeting_transcript, list_recordings, get_recording — plus four prompts. The full schema is in the MCP integration doc.
Can my agent act on transcript chunks in real time?
Yes. The pattern: agent calls request_meeting_bot, polls get_meeting_transcript every few seconds, reasons about the latest segments, and triggers its own tools (Slack message, CRM update, follow-up email) when the transcript matches a rule.
How is this different from giving Claude a generic transcription API?
Two things. (1) MCP gives the agent a typed tool surface — Claude understands the shape of meetings and transcripts natively, no prompt-engineering the API. (2) It works with the platform mix you actually use: drop bots into Meet, Teams, AND Zoom from the same conversation, without rewriting auth flows for each.
Does the MCP server work with self-hosted Vexa?
Yes — point your MCP host at your own gateway's /mcp endpoint instead of api.cloud.vexa.ai. The MCP service is a thin layer over the same REST API, so a self-hosted instance gives you fully on-prem agent-to-meeting access.
Common patterns
What developers build with Vexa MCP
Three agent shapes we see repeatedly across MCP + Vexa builders.
Real-time meeting summarizer agent
Scenario
You want an agent that joins your meetings, summarizes them every 5 minutes, and posts the running summary to a Slack channel — without you running a separate script.
What you get
In Claude: "drop a bot into <URL>, summarize every 5 minutes, post to #sales-calls". Claude orchestrates the bot, reads transcripts, summarizes, and uses its Slack tool to post. Stop the bot with a single prompt.
CRM-enrichment agent triggered by calendar
Scenario
You want an agent that watches your calendar, joins customer calls automatically, transcribes them, and updates the CRM record with action items + sentiment.
What you get
Custom MCP agent: Calendar tool + Vexa tool + HubSpot tool. Calendar trigger → Vexa request_meeting_bot → get_meeting_transcript loop → reason over segments → HubSpot update. Build it once in your agent host, run forever.
Sales-coaching agent reviewing recordings
Scenario
Sales managers want an agent that reviews every rep's calls and writes a coaching note pointing at specific transcript moments.
What you get
Agent runs get_meeting_transcript on completed meetings, finds discovery questions / objections / next steps, and writes a structured coaching review back to your tool of choice (Notion, Linear, Slack DM).
Quickstart
Connect Claude to live meetings in 5 minutes
One command in Claude Code, or one connector URL in Claude Desktop. Your agent gets request_meeting_bot, get_meeting_transcript, list_meetings and six more tools immediately.
- 1
Get a Vexa API key
Sign in at vexa.ai/signin with Google and copy your key from your account page — free credit, no card. Or mint one locally on your self-hosted gateway.
- 2
Add the MCP server
Run the command on the right in Claude Code, or add https://api.cloud.vexa.ai/mcp as a custom connector in Claude Desktop. Nothing to install — it is a remote server, not an npm package.
- 3
Restart your MCP host
Restart your MCP host. The vexa tools (request_meeting_bot, get_meeting_transcript, list_meetings, and six more) appear in the tool list.
- 4
Try it
Ask Claude: "Drop a bot into <Meet/Teams/Zoom URL> and summarize what they discuss." The agent calls request_meeting_bot, then polls get_meeting_transcript.
- 5
Build the agent
Compose vexa tools with your other MCP tools (Slack, Notion, Linear, CRM) to build a meeting-aware agent in your stack.
How it works under the hood
MCP architecture
Vexa MCP is a REMOTE MCP server over streamable HTTP — there is nothing to install and no subprocess. Your host (Claude Code, Claude Desktop, ChatGPT, a custom agent) connects to https://api.cloud.vexa.ai/mcp and authenticates with your Vexa API key; self-hosters point at their own gateway instead. No audio data passes through the MCP host — it stays in the Vexa pipeline (bot pod → Whisper → transcript store).
Pipeline
Documentation
Go deeper on MCP
Tool schemas, transcript chunk format, interactive-bot extensions, self-host setup.
MCP integration guide
Install, configure, and use the Vexa MCP server with Claude Desktop and other MCP hosts.
Interactive bots
Turn the Vexa bot into an interactive participant — speak, chat, share visuals from your agent.
Bots API reference
The underlying REST API the MCP server wraps.
Transcripts API
get_meeting_transcript internals — segment format, polling cadence, language detection.
Authentication
Scope agent access to a single meeting, or to your whole tenant.
Self-host deployment
Run both the MCP server and Vexa fully in your VPC for sovereign agent stacks.
Full reference at docs.vexa.ai.
Related reading
Keep going on MCP
Tutorials, deep-dives, and adjacent surfaces from across the Vexa substrate.
Claude Desktop MCP setup — real-time Google Meet transcripts
The canonical tutorial: add Vexa MCP to Claude Desktop, ask Claude mid-meeting.
Coaching & call review use case
Build an MCP-powered coaching agent that scores calls in real time.
n8n integration
Same goal (orchestration over Vexa) via n8n if your stack already lives there.
A seat, an hour,
or free
A seat for your agent. An hour for what you build. Free when you run it yourself.
Apache 2.0 licensed open source · self-host forever
Best for teams who need full control
Real-time meeting data, directly in your agent (MCP)
The API · +$0.20/hr transcription · $0.50/hr all-in
Best for builders — pay as you go
Vexa Delivery: self-hosted, self-updating, in your own cloud
Best for teams running Vexa in production
Frequently asked questions
What's included in Build on Vexa?
The base rate covers bot infrastructure: audio capture, webhooks, and 12 months of audio storage across Google Meet, Microsoft Teams, and Zoom. Add real-time transcription for +$0.20/hr, so $0.50/hr all-in.
Which plan is right for me?
Two tracks, by the door you come in through. Individual ($12/seat/mo) is for one person and their agents—MCP to connect agents, one concurrent bot, unlimited meetings. Build on Vexa ($0.30/bot-hr) is for products built on the API—pay as you go, unlimited bots. Self-host for free from GitHub if you need full data control, and talk to us for Enterprise.
How does the free credit work?
Every new account gets $2 in free bot credit—no credit card required. That covers ~6 hours of bot time at $0.30/hr. Top-ups start at $10 — Stripe charges a fixed fee per payment, so smaller amounts would mostly be fees; automatic top-ups follow the same minimum. All features are available: audio capture, transcription, real-time data, and full API access.
Can I self-host Vexa?
Yes. Vexa is Apache 2.0 licensed and fully self-hostable. Deploy with Docker Compose, Kubernetes, or OpenShift on your own infrastructure with complete data sovereignty.
How does Vexa compare to Recall.ai?
Vexa is open source, self-hostable, and up to 40% cheaper. Bot: $0.30/hr vs Recall.ai ~$0.50/hr. Transcription add-on: $0.20/hr vs ~$0.15/hr. See our detailed comparison.
What's the Individual plan for?
Individual is the seat: $12 per seat per month for one person and the agents they connect over MCP. One concurrent bot, unlimited meetings, real-time transcription, recording, storage, REST API, WebSockets, and the UI Dashboard. It supports Google Meet, Microsoft Teams and Zoom.
Ready when you are
Connect your agent.
Live meeting context in 5 minutes.
No credit card, no SDK install — just an MCP server entry. Add it to Claude Desktop, restart, ask your agent to drop a bot into a meeting.
Apache 2.0 · Self-host or cloud · No credit card required