Up to 30% cheaper than Recall.ai — simple, agent-friendly API. See pricing
Zoom Google Meet Microsoft Teams

Simple, agent-friendly API. 3 endpoints. That's it.

Capture audio, transcribe, and stream meeting data from Zoom, Google Meet, and Teams. Built for AI agents to use programmatically. Pay-as-you-go, no subscriptions, no lock-in.

terminal
curl -X POST https://api.firstcall.dev/v1/bots \
  -H "Authorization: Bearer ak_live_xxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "meet_url": "https://meet.google.com/abc-def-ghi",
    "bot_name": "AI Assistant",
    "mode": "audio-ws",
    "transcription": true
  }'

# Response: { "bot_id": "bot-xxx", "ws_url": "wss://...", "created_at": "..." }
# Connect WebSocket: wss://api.firstcall.dev/v1/bots/bot-xxx/ws?api_key=ak_live_xxx

Why teams choose us

Lower cost. Agent-first design. Simple pay-as-you-go pricing.

$
~30%

Lower Cost

$0.35/hour bot compute + $0.12/hour optional transcription. Combined $0.47/hr vs their $0.65/hr. No hidden fees. No minimum commitments. Prorated to the second.

Built for AI Agents

Real-time audio streaming via WebSocket, webhook-driven events, and programmatic bot control. Your agents can join meetings, listen, and process audio in real-time.

Dead Simple API

3 REST endpoints: create, status, stop. WebSocket for real-time data. Webhooks for async events. No complex setup, no lock-in.

Up and running in minutes

One POST request. Real-time meeting data back via WebSocket and webhooks.

1

Send a meeting URL

Pass any Zoom, Google Meet, or Teams link to our API. One POST request. That's all it takes.

2

We handle the rest

Our bot joins in under 3 seconds, captures audio, takes periodic screenshots, and optionally transcribes in real-time.

3

Get structured data

Receive audio streams, transcripts, and meeting events via WebSocket or webhooks. JSON-first.

Works where your team meets

Google Meet Full support
Microsoft Teams Enterprise-grade integration
Zoom Native bot with all features

What's under the hood

Audio capture, transcription, real-time streaming, and agent tooling — all through a single API.

Core

Meeting Bot API

Send a meeting link, get a bot that captures audio, takes periodic screenshots, and delivers transcripts and participant metadata.

  • Real-time audio streaming with sub-200ms latency
  • Automatic speaker identification and labeling
  • Periodic screenshot capture
  • Webhook-based event delivery
  • Bot customization: name, avatar, join behavior
Agents

AI Agent Mode

Feed real-time meeting audio to your LLM. Your agent can join meetings, listen, and process speech as it happens.

  • Real-time audio input stream for agent consumption
  • WebSocket delivery of audio chunks and transcripts
  • Context-aware webhooks with speaker and topic data
  • Custom bot branding via webpage modes (modes 3-4)
  • Multi-agent support in the same meeting
Transcription

Real-Time Transcription

Built-in transcription at $0.12/hour with real-time delivery. Enable with a single flag.

  • Real-time transcript streaming via WebSocket
  • Speaker diarization with participant names
  • Post-call polished transcripts with timestamps
  • Optional — only pay when you enable it
Developer

Developer-Friendly

Simple REST API with 3 endpoints. WebSocket for real-time data. Webhooks with signature verification for async events.

  • 3 REST endpoints: create bot, get status, stop bot
  • WebSocket for real-time audio and transcript streams
  • Webhook signature verification (HMAC-SHA256)
  • Full API docs at /docs/rest-api.md

Three ways to get meeting data

REST API to control bots. WebSocket for real-time data. Webhooks for async events.

REST API

3 endpoints. Create a bot, check status, stop it. That's the entire API.

  • POST /v1/bots — Create & join a meeting
  • GET /v1/bots/:id — Check bot status
  • DELETE /v1/bots/:id — Leave meeting
REST API reference →

WebSocket

Connect to the bot's WebSocket for real-time meeting data as it happens.

  • transcript.final — Transcripts with speaker name
  • meeting.participant_joined — Who joined/left
  • audio.chunk — Raw PCM audio stream
  • screenshot.result — Meeting screenshots
WebSocket API reference →

Webhooks

Get notified when bots change state. HMAC-SHA256 signed, with automatic retries.

  • bot.created — Bot was created
  • bot.status_changed — Bot joined, ready, etc.
  • bot.ended — Meeting over, with cost breakdown
Webhook API reference →

AI agents: read our complete API at firstcall.dev/llms.txt

Pay-as-you-go pricing

No subscriptions. No minimums. Add credits anytime. Up to 30% less than Recall.ai.

Pay As You Go
Bot Compute
$0.35/hr
Audio capture + screenshots
Transcription
$0.12/hr
Optional — enable per bot
  • 6 free hours (360 minutes) to start — no credit card required
  • 1 concurrent bot on free tier, unlimited for paid
  • All platforms supported (Zoom, Google Meet, Teams)
  • Real-time audio streaming via WebSocket
  • 7-day session retention
  • Prorated to the second — no rounding
  • Add credits anytime, no subscription
Get Started Free

firstcall.dev vs Recall.ai

Service firstcall.dev Recall.ai
Bot compute (per hour)$0.35$0.50
Transcription (per hour)$0.12$0.15
Combined (per hour)$0.47$0.65
Free tier6 hours (360 min)5 hours
Session retention7 days7 days
Pricing modelPay-as-you-goTiered plans
AI Agent ModeNativeLimited

Start building in 5 minutes

Sign up, get an API key, send your first bot.

1

Sign up and get your API key

Create your firstcall.dev account. Free tier includes 6 hours (360 minutes) of bot compute. No credit card needed.

2

Create your first bot

POST to /v1/bots with a meeting URL, bot name, and mode. Get back a bot ID and WebSocket URL.

3

Connect via WebSocket

Open a WebSocket connection to receive real-time audio, transcripts, and meeting events as they happen.

4

Stop when done

POST to /v1/bots/:bot_id/stop to remove the bot from the meeting. You only pay for time used.

Built for AI agents to use programmatically

3 endpoints your agent can call. WebSocket for real-time meeting data.

📡

Real-time Streaming

WebSocket streams deliver audio chunks and live transcription with sub-200ms latency. Process speech as it happens.

🔊

Audio Capture

Raw PCM audio streamed via WebSocket. Feed directly into your speech models or processing pipeline.

🧠

Context-Aware Webhooks

Structured payloads include who's speaking, participant data, and meeting events. Context, not raw data.

🎨

Custom Bot Branding

Use webpage modes (3-4) to display custom branded content in the meeting. Your agent, your brand.

👥

Multi-Agent Support

Deploy multiple agents to one meeting, each with different roles. Note-taker, fact-checker, decision tracker in parallel.

📸

Periodic Screenshots

Automatic periodic screenshots of the meeting screen. Capture shared content and visual context alongside audio.

Agent use cases

Sales AI Agent

Joins sales calls, captures real-time transcription, auto-updates CRM with notes, follow-ups, and deal stage changes.

Interview Agent

Listens to interviews, generates structured scorecards, flags inconsistencies, suggests follow-up questions.

Customer Success Agent

Monitors calls for churn signals and escalation triggers. Automatically creates tickets and alerts the team.

Compliance Agent

Monitors calls in regulated industries for compliance keywords and generates audit trails from transcripts.

Frequently asked questions

What is firstcall.dev? +

firstcall.dev is a meeting bot API that lets developers programmatically capture audio, transcribe, and stream data from Zoom, Google Meet, and Microsoft Teams meetings. It's designed for both traditional meeting intelligence apps and autonomous AI agents.

How is firstcall.dev different from Recall.ai? +

We focus on simplicity and cost: 3 REST endpoints, WebSocket for real-time data, up to 30% cheaper ($0.47/hr combined vs $0.65/hr). No complex tiered plans — just pay-as-you-go with no subscriptions or lock-in.

Which meeting platforms do you support? +

We currently support Google Meet, Microsoft Teams, and Zoom.

Is there a free tier? +

Yes. Every account starts with 6 free hours (360 minutes) of bot compute. No credit card required. Free tier includes 1 concurrent bot. After that, add credits anytime at $0.35/hour for bot compute and $0.12/hour for transcription.

What data do I get from meetings? +

You get real-time audio streams (PCM 16-bit), optional transcription with speaker labels, periodic screenshots, participant join/leave events, and meeting lifecycle webhooks. All delivered via WebSocket and webhooks in JSON format.

How does pricing work? +

Pay-as-you-go, prorated to the second. Bot compute costs $0.35/hour. Transcription is optional at $0.12/hour. 7-day session retention included. No seat-based pricing, no monthly minimums, no subscriptions. Add credits anytime.

How many concurrent bots can I run? +

Free tier allows 1 concurrent bot. Once you add credits (paid tier), there is no limit on concurrent bots.

Try it free

6 free hours (360 minutes) included. No credit card required.