Elevenlabs

ElevenLabs

Teler streams call audio between a live phone call and your ElevenLabs agent (ElevenLabs Agents, formerly Conversational AI), so callers talk to the agent you built in ElevenLabs.

Category

Overview

ElevenLabs Conversational AI runs your agent (voice, persona, knowledge, tools) and exposes a WebSocket per session. The bridge proxies Teler call audio into that WebSocket and relays the agent’s response back.

Unlike the OpenAI and Gemini bridges, this one is purely an audio pipe – your agent’s voice, tools, prompt, and knowledge base are all configured in the ElevenLabs dashboard, not in the bridge itself. The bridge holds two WebSockets, one to Teler and one to ElevenLabs, and pumps audio between them.

Which should I use?

  • SIP trunk: Set up in the Teler and ElevenLabs dashboards, with no code or server. Calls work in both directions. → See the SIP trunk setup
  • Bridge (Media Streaming): You run a small server in Docker that passes call audio between Teler and your ElevenLabs agent.

Key Features

  • Reference implementations in Python and Node – Clone teler-elevenlabs-bridge (Python) or teler-elevenlabs-node-bridge (Node) and get a working bridge in minutes.
  • One-command Docker setup – Clone the repo, fill in your environment variables, and start with a single command.
  • Agent configuration stays in ElevenLabs – voice, tools, prompt, and knowledge base are all managed in the ElevenLabs dashboard, so non-engineers can iterate on the agent without touching the bridge.

How to Integrate

Prerequisites

  • An ElevenLabs account with a Conversational AI agent already created
  • A Teler account with an active API key
  • One Teler phone number on your account
  • An ngrok auth token (for local development)
  • Docker and Docker Compose

Step 1: Clone and configure

git clone https://github.com/frejun-tech/teler-elevenlabs-bridge
cd teler-elevenlabs-bridge
cp .env.example .env
# fill ELEVENLABS_WEBSOCKET_URL, TELER_API_KEY, NGROK_AUTHTOKEN
docker compose up --build

ELEVENLABS_WEBSOCKET_URL is your agent’s WebSocket URL, e.g. wss://api.elevenlabs.io/v1/convai/conversation?agent_id=<id>.

Step 2: Initiate a call

curl -X POST http://your-domain.com/api/v1/calls/initiate-call \
  -H "Content-Type: application/json" \
  -d '{
    "from_number": "+91XXXXXXXXXX",
    "to_number":   "+91XXXXXXXXXX"
  }'

Integration at a Glance

RequirementsElevenLabs account with a Conversational AI agent created, Teler account with API key and phone number, Docker + Docker Compose, ngrok auth token (local dev)
Supportsupport@frejun.com, or file an issue directly on the GitHub repo for bridge-specific bugs

Frequently Asked Questions

Have more questions?
Reach out to us at support@frejun.com.

The agent answers, but the voice is wrong - why?

Check for the wrong agent_id in your WebSocket URL - the agent must exist in your ElevenLabs account.

The bridge is probably unreachable from the public internet. Check that ngrok is running and verify SERVER_DOMAIN.

Likely a sample rate mismatch - try setting ELEVENLABS_SAMPLE_RATE=8k.

Usually background noise or low input volume - test with a clean audio source first.

In the ElevenLabs dashboard, not in the bridge. The bridge only handles the audio connection between Teler and your agent.

Yes, with Teler in two ways: a SIP trunk set up in the dashboards, or a bridge you run yourself that streams call audio to your agent.

Use the SIP trunk for a dashboard setup with no code. Use the bridge if you’d rather run the connection yourself, as a Docker container, in Python or Node.

No. Voice, prompt, tools and knowledge base stay in ElevenLabs. The bridge only carries the call audio.

Get started with Teler today

Real-time voice APIs for Al agents. WebSocket streaming.
SIP trunking, global numbers. Ship voice applications in hours.