Google Meet Bot API — Transcription & Meeting Intelligence

Programmatic Google Meet bots that join calls, capture speaker-diarized transcripts, and deliver structured meeting data via REST API — built as an Apify Actor for developers who want meeting intelligence without building browser automation from scratch.

Send a bot to any Google Meet. Get diarized transcripts and meeting metadata — all through Apify's production-grade API, webhooks, and integrations.

---

Why this over Recall.ai, Vexa, or rolling your own?

	Lexis Meet Agent	Typical meeting-bot APIs
API platform	Full Apify REST API — runs, datasets, KV store, webhooks, schedules	Custom REST + WebSocket
Transcripts	Full diarized transcript written to Dataset and KV store at meeting end	Real-time WebSocket / webhook
Speaker diarization	Native Google Meet captions with per-speaker attribution	STT diarization
Audio / video recording	Not included (captions-only)	MP4, separate audio streams
Source code	Open Actor — fork, self-host, audit	Closed / partial open source
Scaling	Apify cloud — concurrent runs, retries, monitoring	Managed infra
Integrations	n8n, Zapier, Make, webhooks, any HTTP client	Platform-specific
Join model	Guest bot (host admits) or authenticated Google session	Often no host permission required

Best fit: engineering teams building meeting intelligence, AI notetakers, sales call logging, interview transcription, compliance workflows, or agentic apps — who want API-first control and Apify's developer ecosystem instead of a closed meeting-BaaS.

Compare: Recall.ai Google Meet Bot API · Vexa Meeting Transcription API

---

What you get

Google Meet Bot API primitives

• Automatic join & leave — bot joins via meeting URL, stays for the call, exits on alone-timeout, max duration, removal, or exit phrases
• Custom bot identity — set display name per meeting (botName)
• Transcription — live captions observed via DOM MutationObserver; full diarized transcript written at meeting end
• Speaker-diarized transcripts — each segment includes speaker and text
• Meeting metadata — run status, segment counts, end reason, meeting URL in status.json and transcript.json
• Post-meeting output — full TRANSCRIPT.md and structured transcript.json
• Multi-language captions — configure caption language (en, es, de, …)
• Agent-ready data — feed transcripts into LLMs, CRMs, ticketing, or RAG pipelines

Built for AI agents & automation

• REST API for every operation — start bots, read final transcripts and artifacts
• Webhooks on run finish — trigger n8n, Zapier, or your backend when a meeting ends
• Schedules — calendar-driven bot deployment via Apify Schedules
• Concurrent bots — run unlimited parallel meetings (Apify platform limits apply)
• Observable runs — live logs, screenshots, and run history in Apify Console

---

Two API calls to meeting intelligence

1. Start a Google Meet bot

curl -X POST "https://api.apify.com/v2/acts/YOUR_USERNAME~google-meet-transcription-bot/runs?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "meetingUrl": "https://meet.google.com/abc-defg-hij",
    "botName": "Acme Notetaker",
    "captionLanguage": "en",
    "aloneTimeoutSecs": 120
  }'

2. Get the transcript (after the meeting ends)

Transcript segments (available once the run completes):

# Structured JSON
curl "https://api.apify.com/v2/key-value-stores/STORE_ID/records/transcript.json?token=YOUR_APIFY_TOKEN"

# Human-readable Markdown
curl "https://api.apify.com/v2/key-value-stores/STORE_ID/records/TRANSCRIPT.md?token=YOUR_APIFY_TOKEN"

# Run status (joining → captions_active → ended)
curl "https://api.apify.com/v2/key-value-stores/STORE_ID/records/status.json?token=YOUR_APIFY_TOKEN"

TypeScript / JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

// Join the meeting
const run = await client.actor('YOUR_USERNAME/google-meet-transcription-bot').call({
    meetingUrl: 'https://meet.google.com/abc-defg-hij',
    botName: 'Vexa-style Notetaker',
});

// Diarized transcript segments
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);
// [{ speaker: "Alice", text: "..." }, ...]

// Full transcript artifact
const store = client.keyValueStore(run.defaultKeyValueStoreId);
const transcript = await store.getRecord('transcript.json');

Python

from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])

run = client.actor("YOUR_USERNAME/google-meet-transcription-bot").call(run_input={
    "meetingUrl": "https://meet.google.com/abc-defg-hij",
    "botName": "Meeting Bot",
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["speaker"], item["text"])

Webhook on meeting end

const run = await client.actor('YOUR_USERNAME/google-meet-transcription-bot').call(
    { meetingUrl: 'https://meet.google.com/abc-defg-hij' },
    {
        webhooks: [
            {
                eventTypes: ['ACTOR.RUN.SUCCEEDED', 'ACTOR.RUN.FAILED'],
                requestUrl: 'https://your-app.com/webhooks/meeting-ended',
            },
        ],
    },
);

Or configure a permanent webhook on the Actor in Apify Console → your Actor → Integrations → Webhooks.

Your webhook receives the run payload with links to the dataset and key-value store — fetch transcripts and push to Salesforce, HubSpot, Linear, Slack, or your vector DB.

---

How it works

Meeting URL  →  Apify API  →  Bot joins Meet  →  Captions ON  →  Bot leaves  →  Transcript artifacts

1. Launch — Camoufox (anti-fingerprint Firefox) via Playwright on Apify infrastructure
2. Join — guest join with your botName, mic/camera off, keyboard-driven lobby navigation
3. Admit — waits for host admission (configurable timeout)
4. Transcribe — enables Google Meet live captions, observes caption DOM with MutationObserver
5. Diarize — accumulates per-speaker turns; speaker change triggers segment commit
6. End — leaves on alone-timeout, max duration, removal, or configurable exit phrases
7. Deliver — writes full TRANSCRIPT.md, transcript.json, and status.json to the KV store; pushes segments to the Dataset

No audio capture. No speech-to-text API. Transcription quality follows Google Meet captions for the selected language.

---

Input

Field	Required	Default	Description
meetingUrl	Yes	—	Google Meet link (https://meet.google.com/...)
botName	No	Notetaker	Bot display name shown to participants
maxDurationSecs	No	7200	Max time in meeting (0 = unlimited)
captionLanguage	No	en	Caption language code (en, es, de, …)
aloneTimeoutSecs	No	5	Leave when bot is alone this long
admissionTimeoutSecs	No	600	Wait for host to admit the bot
storageStatePath	No	—	Playwright storage state from a signed-in Google account (more reliable than guest)

---

Output

Dataset — transcript segments

One row per speaker turn, written to the Dataset when the meeting ends:

{
    "speaker": "Alice Chen",
    "text": "We need to finalize the API contracts by Friday."
}

Fetch GET /v2/actor-runs/{runId}/dataset/items after the run completes.

Key-Value Store — meeting artifacts

Key	Description
transcript.json	Full structured transcript + metadata
TRANSCRIPT.md	Human-readable transcript
status.json	Live run status (joining → captions_active → ended)
SELECTOR_PROBE.json	Debug: caption selector health

Status lifecycle

starting → joining → admitted → captions_active → ended

Terminal states: blocked_guest, admission_timeout, error

---

Use cases

• Sales & revenue — auto-log discovery calls, demo notes, competitor mentions to CRM
• Engineering — standup / retro transcripts for AI coding agents and ticket creation
• HR & recruiting — interview transcription with speaker attribution
• Compliance — meeting recordkeeping where caption-based capture is acceptable
• AI notetakers — build your own Otter/Fireflies alternative on Apify
• Workflow automation — n8n/Zapier triggers on ACTOR.RUN.SUCCEEDED

---

Requirements & limitations

Works with

• Google Meet on all common Workspace tiers (guest join where org policy allows)
• Free Google accounts when guest access is permitted
• Multi-language live captions supported by Google Meet

Limitations

• Host must admit the guest bot (unless using an authenticated storageStatePath session)
• Some organizations block guest joins — run exits with blocked_guest
• Captions-only — no audio recording, video recording, screenshare capture, or chat messages
• Google Meet only — Microsoft Teams and Zoom not supported in this Actor
• Google Meet's DOM is not a public API; selectors may need updates when Meet changes
• Caption quality varies by language, accent, and Meet's own STT
• Real-time delivery is via Dataset polling (sub-second segment finalization; not a dedicated WebSocket)

Consent

This Actor captures meeting captions, which may constitute recording in some jurisdictions. You are responsible for obtaining consent from meeting participants before use.

---

Architecture

Layer	Technology
Platform	Apify Actors
Browser	Camoufox + Playwright (Firefox)
Interaction	Ghost cursor, semantic locators, keyboard shortcuts
Transcription	Google Meet live captions (DOM scrape)

---

FAQ

Does this work as a Google Meet transcription API? Yes. Start a run with a Meet URL; retrieve diarized transcript segments from the Dataset API and full artifacts once the run completes.

Do I need the host's permission? For guest join, the host (or someone with admit rights) must let the bot into the meeting. For stricter orgs, use a signed-in Google session via storageStatePath.

Can I get real-time transcripts? Not currently — segments are written to the Dataset when the meeting ends. Use the status.json key-value record to monitor the bot's state (joining → captions_active → ended) and fetch the transcript after the run completes.

How does this compare to Recall.ai? Recall.ai offers audio/video capture, wider platform support, and often no-admit joins. This Actor is caption-based (simpler compliance surface), open source, and runs on Apify's API with webhooks, schedules, and integrations you may already use.

How does this compare to Vexa? Vexa is open-source meeting-bot infrastructure with WebSockets and self-hosting. This Actor gives you a deploy-ready Google Meet bot on Apify with the same "send URL → get transcript" developer experience, plus Apify's managed scaling and ecosystem.

Can I run multiple bots at once? Yes. Each meeting is a separate Apify Actor run. Scale concurrent bots via Apify platform limits and billing.

Can I schedule bots for calendar meetings? Use Apify Schedules or trigger runs from your calendar integration (Google Calendar → webhook → Apify API).

Is there speaker diarization? Yes — speaker names come from Google Meet's live caption UI, attributed per segment.

---

Google Meet Bot API · Meeting Transcription API · Real-Time Transcripts · Speaker Diarization · Meeting Intelligence · Apify Actor

---

👀 p.s.

Got feedback or need an extension?

Lexis Solutions is a certified Apify Partner. We can help you with custom solutions or data extraction projects.

Contact us over Email or LinkedIn

Support Our Work 💝

If you're happy with our work and scrapers, you're welcome to leave us a company review here and leave a review for the scrapers you're subscribed to. It will take you less than a minute but it will mean a lot to us!