Vision Agents Plugin: Tencent TRTC Edge

Tencent TRTC (Real-Time Communication) edge transport for Vision Agents. Lets an agent join a Tencent TRTC room and exchange audio/video with participants using the Tencent LiteAV SDK Python bindings.

Quickstart

Talk to an agent in 5 minutes. You'll need Docker, an .env with TENCENT_SDK_APP_ID, TENCENT_SDK_SECRET_KEY, OPENAI_API_KEY, and ELEVEN_API_KEY at the repo root, and a working microphone in a Chromium-based browser. The example pairs Tencent TRTC with OpenAI (LLM) and ElevenLabs (STT + TTS) — chosen because Gemini and Cartesia aren't reachable from mainland China, which is the natural deployment target for this edge.

1. Open Tencent's hosted TRTC Web SDK quick demo: https://web.sdk.qcloud.com/trtc/webrtc/v5/demo/quick-demo-js/index.html

   • Paste your TENCENT_SDK_APP_ID into SDKAppID and TENCENT_SDK_SECRET_KEY into SDKSecretKey — the page generates UserSig client-side.
   • Leave the auto-generated UserID and RoomID(String) as is.
   • Click Enter Room → demo log should print 🟩 [user_***] enterRoom.
   • Click Start Local Video — this also publishes the mic.

2. Copy the RoomID(String) from the demo form and launch the agent with it:

   cd plugins/tencent
   TENCENT_TEST_ROOM_ID=<paste-room-id-here> docker compose run --rm tencent-agent
   

   On first run this builds the image and resolves the workspace; subsequent runs are fast. When the agent joins you'll see Tencent TRTC OnRemoteUserEnterRoom: <userId> matching the UserID shown in the demo.

3. Talk. The flow is browser → Tencent → STT (ElevenLabs) → LLM (OpenAI) → TTS (ElevenLabs) → Tencent → browser. Confirm each leg in the agent log:

   • 🎤 [Transcript Complete]: … — STT got your speech.
   • 🤖 [LLM response final]: … — LLM produced a reply.
   • Reply plays back in the browser.

Usage in your own code

from vision_agents.core import Agent, User
from vision_agents.plugins import tencent

agent_user = User(id="agent-1", name="Agent")
edge = tencent.Edge(
    sdk_app_id=YOUR_SDK_APP_ID,
    user_sig=USER_SIG,  # or key=KEY to generate per join
)

agent = Agent(
    edge=edge,
    agent_user=agent_user,
    llm=...,
    # stt, tts, etc.
)

call = await edge.create_call(call_id=str(ROOM_ID), room_id=ROOM_ID)
await agent.join(call)
await agent.finish()

Install

uv add vision-agents-plugins-tencent

On Linux this pulls liteav from PyPI. On macOS the liteav dependency is skipped via a platform marker, so the package installs but tencent.Edge() raises at runtime — use the Docker setup from the Quickstart.

Configuration

tencent.Edge(...) parameters:

• sdk_app_id (int): Tencent TRTC SDK App ID. Falls back to TENCENT_SDK_APP_ID env var.
• user_sig (str): User signature for the agent user.
• key (str): Optional. If set and user_sig is not, the plugin generates user_sig via TLSSigAPIv2. Falls back to TENCENT_SDK_SECRET_KEY env var.
• video_fps (int): Outgoing video frame rate.

Environment variables:

• TENCENT_SDK_APP_ID, TENCENT_SDK_SECRET_KEY — credentials.
• TENCENT_TRTC_SCENE — one of auto (default), videocall, call, record.
• TENCENT_TEST_ROOM_ID — interpolated by docker-compose.yml into the example runner's --call-id flag.

Platform support

The underlying liteav package ships only manylinux wheels — Linux x86_64 / aarch64 only. macOS and Windows must run the agent inside a Linux container (see Quickstart). User sigs are needed for room entry; either generate them with TLSSigAPIv2 or pass key= and let the plugin sign per join.