CHECKED OCT 10, 2026
GA Agora / TEN community

TEN Framework

Open-source framework for realtime multimodal conversational agents backed by Agora, with a visual graph designer (TMAN Designer), agent examples, and the TEN VAD and TEN Turn Detection models.

Est. per minuten/a

Overview

Best for: Teams on Agora who want to self-host and customise the same stack Agora's engine is built on, or who need the TEN VAD/turn models.

At a glance

Platform $/min$0
All-in priceNo
Entry plan $/mo$0
Free tierYes
Phone numbersNo
Own LLMYes
Own keysYes
No-codeYes
Smart turnsYes
WebRTCYes
Phone / SIPYes
Self-hostYes
Open weightsYes

Open source (Apache 2.0 with additional restrictions). Turn detection model covers English and Chinese only. Visual graph designer (TMAN). SIP via example extension, no numbers.

Audio in

Agora RTC by default; other extensions possible

Audio out

Same

Languages

Depends on providers; turn detection supports English and Chinese.

Latency

No verified number.

Regions

Self-hosted.

Compliance

N/A (self-hosted).

Features

  • visual graph designer
  • turn detection
  • VAD
  • multi-provider plug-ins (extensions)
  • avatars
  • SIP call example

Pricing

WhatPriceUnitNotes
Framework$0licenseYou pay Agora RTC (or other transport) and every model provider separately.
How the per-minute estimate was worked out

Depends entirely on chosen transport and providers; no vendor figure.

Free tier: Open source.

Source: github.com

Setup

  1. Clone github.com/TEN-framework/ten-framework and follow the ai_agents README (Docker, Node.js 18 prerequisite).
  2. Fill in .env with Agora App ID and provider keys.
  3. Start the dev stack, open TMAN Designer at localhost:49483, pick an example graph (voice assistant, SIP call) and edit extension properties.

Endpoint

Self-hosted

Authentication

Provider keys in .env

Quick start python

# TEN is configured as a graph (JSON) of extensions rather than a short script.
# Start from an example in the repo:
#   git clone https://github.com/TEN-framework/ten-framework
#   cd ten-framework/ai_agents   # follow README: copy .env.example to .env, add keys
#   # start the dev containers as the README describes, then open
#   # http://localhost:49483  (TMAN Designer) and select the voice-assistant graph

Written from the current docs. Check the vendor's SDK version before you ship.

Warnings

License has extra restrictions

Both the framework and the turn-detection model are 'Apache 2.0 with additional restrictions'. Read LICENSE before commercial use.

Smaller ecosystem

Fewer integrations, docs and community answers than Pipecat or LiveKit Agents.

No managed hosting from the project

Hosted use is via Agora's Conversational AI Engine (priced separately).

Releases not clearly versioned

The GitHub releases section was empty when checked; pin a commit.

Plus 12 warnings that apply to all platforms and telephony APIs. See category warnings.

Limits

  • You operate everything: scaling, monitoring, TURN

Models and products

NameStatusNotes
TEN FrameworkGAApache 2.0 with additional restrictions (packages directory is plain Apache 2.0).
TEN VADGALightweight low-latency streaming voice activity detector.
TEN Turn DetectionGASemantic end-of-turn model for full-duplex dialogue, English and Chinese; Apache 2.0 with additional restrictions.

Docs and sources

Docs

Sources used

Not fully verified

Exact startup commands (README changes often); latest version.

Similar platforms APIs

Vonage (Ericsson)Vonage Voice/Video API AI connectors (Audio Connector) + AI Studion/a0 high-severity warningsOpenAIOpenAI Agents SDK (voice / realtime agents)n/a0 high-severity warningsVercelVercel AI SDK realtime (via AI Gateway)n/a0 high-severity warningsVoiceflowVoiceflow (voice agents)n/a0 high-severity warnings
Spotted a wrong price or a dead link?