Turn your meetings into real-time transcripts, concise summaries, and actionable next steps. Telnyx meeting bots join automatically, capture conversations as they happen, and organize the key takeaways when the call ends.

Create a session, poll for status, fetch the transcript, and tell the bot to speak.
Create, poll, transcribe, speak
# 1. Create a meeting session
curl -X POST https://api.telnyx.com/v2/meeting_sessions \
-H "Authorization: Bearer $TELNYX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"meeting_url": "https://meet.google.com/abc-defg-hij",
"bot_name": "Standup Bot",
"summarize_on_end": true
}'
# 2. Poll session status until active
curl https://api.telnyx.com/v2/meeting_sessions/$SESSION_ID \
-H "Authorization: Bearer $TELNYX_API_KEY"
# 3. Fetch the live transcript
curl "https://api.telnyx.com/v2/meeting_sessions/$SESSION_ID/transcript?after=0&limit=100" \
-H "Authorization: Bearer $TELNYX_API_KEY"
# 4. Make the bot speak
curl -X POST https://api.telnyx.com/v2/meeting_sessions/$SESSION_ID/actions/speak \
-H "Authorization: Bearer $TELNYX_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "text": "I have the action items from last week." }'Same flow with requests
import os, time, requests
API = "https://api.telnyx.com/v2"
HEADERS = {"Authorization": f"Bearer {os.environ['TELNYX_API_KEY']}", "Content-Type": "application/json"}
# 1. Create a meeting session
session = requests.post(f"{API}/meeting_sessions", headers=HEADERS, json={"meeting_url": "https://meet.google.com/abc-defg-hij", "bot_name": "Standup Bot", "summarize_on_end": True}).json()["data"]
session_id = session["id"]
# 2. Poll until active
while True:
status = requests.get(f"{API}/meeting_sessions/{session_id}", headers=HEADERS).json()["data"]
if status["status"] == "active": break
time.sleep(5)
# 3. Fetch the transcript
transcript = requests.get(f"{API}/meeting_sessions/{session_id}/transcript?after=0&limit=100", headers=HEADERS).json()["data"]
# 4. Make the bot speak
requests.post(f"{API}/meeting_sessions/{session_id}/actions/speak", headers=HEADERS, json={"text": "I have the action items from last week."})Same flow with fetch
const API = "https://api.telnyx.com/v2";
const headers = {Authorization: `Bearer ${process.env.TELNYX_API_KEY}`, "Content-Type": "application/json"};
// 1. Create a meeting session
const session = await fetch(`${API}/meeting_sessions`, {method: "POST", headers, body: JSON.stringify({meeting_url: "https://meet.google.com/abc-defg-hij", bot_name: "Standup Bot", summarize_on_end: true})}).then(r => r.json());
const sessionId = session.data.id;
// 2. Poll until active
let status;
do {
await new Promise(r => setTimeout(r, 5000));
status = await fetch(`${API}/meeting_sessions/${sessionId}`, {headers}).then(r => r.json());
} while (status.data.status !== "active");
// 3. Fetch the transcript
const transcript = await fetch(`${API}/meeting_sessions/${sessionId}/transcript?after=0&limit=100`, {headers}).then(r => r.json());
// 4. Make the bot speak
await fetch(`${API}/meeting_sessions/${sessionId}/actions/speak`, {method: "POST", headers, body: JSON.stringify({text: "I have the action items from last week."})});
Every other notetaker assumes a human clicks join. Meeting API assumes an agent calls an endpoint. Same primitives, same API key, extended into the meeting room.

Live transcripts come from Telnyx STT, the bot speaks through Telnyx TTS, and summaries and action items come from Telnyx Inference. Your agent does not need a separate transcription or model provider.

The assistant answers questions and acts on what is said in the meeting, using the same transcript the API returns to you. You can also connect your own agent through MCP or the REST API.
Each primitive is independently useful and composes with the rest of the Telnyx stack.
Live transcription
Stream the meeting transcript to your agent as it happens, with speaker labels and timestamps, through Telnyx STT.
Programmable speech
Your agent can make the bot say anything at any point in the meeting. Text goes through Telnyx TTS and plays into the meeting audio.
Chat injection
Drop messages into the meeting chat on Zoom, Google Meet, and Teams.
Post-meeting artifacts
When the meeting ends, receive structured summaries and action items generated by Telnyx Inference. summarize_on_end auto-generates a summary from the final transcript.
Webhook-driven lifecycle
Set a webhook_url on session create to receive signed meeting lifecycle events. No polling required, but polling is supported for batch workflows.
Cross-platform support
One API surface for Zoom, Google Meet, and Microsoft Teams. Same request shape, same response shape, same webhook contract.

Programmable voice agents for inbound and outbound calls. Same STT, TTS, and orchestration primitives that power Meeting API.

Open-weight and frontier models on edge-deployed GPUs. Per-token pricing with global availability.

Hosted voices from Telnyx and third-party providers. Use the same voice in your meeting bot, your voice agent, and your IVR.

Real-time and batch transcription with multiple model options. Pick the model that fits your latency, accuracy, and cost targets.
Sign up, grab your key, and run the cURL example above. Meeting API is in beta.
Sign up, grab your key, and run the cURL example above. Meeting API is in beta.
Meeting API is an API-first agent primitive that lets a developer programmatically add a Telnyx-controlled AI participant to a Zoom, Google Meet, or Microsoft Teams meeting. The bot transcribes the meeting live, can speak and chat into it, and returns structured summaries and action items when the meeting ends.
Zoom, Google Meet, and Microsoft Teams. The platform is detected automatically from the meeting URL. Chat works on all three. Recording export works on Google Meet and Teams.
Your agent makes API calls: create a session with a meeting URL, poll for status until the bot is active, fetch the transcript, and optionally tell the bot to speak or send a chat message. Webhooks are available for event-driven workflows.
Meeting API is billed per minute of bot time in the meeting, and that rate includes live transcription. Speech, AI assistants, and summaries and action items are billed separately, and recordings are retained free for 7 days. See the Meeting API pricing page in the developer docs for current rates. Meeting API is in beta.
The bot transcribes through Telnyx STT. Model selection is not currently exposed through the API.
Yes, on Google Meet and Teams. Set recording_export when you create the session and the video is exported to your Telnyx Storage bucket. Recording export is not available for Zoom meetings. Transcription works with or without recording.
The bot can generate summary and action_items artifacts through Telnyx Inference. summarize_on_end auto-generates a summary once the final transcript settles after the meeting ends.
Set a webhook_url on session create to receive signed meeting lifecycle events. Webhooks are off by default.
Transcripts, events, and artifacts are stored in Telnyx Postgres. Recordings are retrievable through the API.
No. Meeting API is in beta. The API surface and feature set may change before general availability.
A standalone meeting bot API gives you a bot in the meeting and a transcript. Telnyx gives you the bot, the transcript, the speech, the summaries, and the same API key you already use for calls, STT, TTS, and inference. No second vendor, no second contract, no stitching a meeting bot API to a separate transcription provider and a separate LLM for summaries. You also control the bot's voice, when it speaks, what it says, when it interrupts, and how artifacts are generated, all through the same API. Unlike traditional meeting bots that only record and transcribe, you can attach a Telnyx AI Assistant that listens to the meeting, reasons about what's being said, and speaks back in real time.