OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Voice Agents

Vision Agents

Vision Agents is an open Python framework for building real-time voice and vision agents from interchangeable AI models, speech components, video providers and processing pipelines. It supports live WebRTC media, model calls, external tools and video analysis, with Stream’s edge network available for low-latency operation.

Independent OpenViST project profile, based on linked sources.

Vision Agents — OpenViST project profile in Voice Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Vision Agents when a project needs a programmable real-time agent that combines voice interaction with live video understanding and replaceable provider integrations.

Who it is for

  • Python developers building conversational voice agents
  • Teams developing agents that interpret live video
  • Prototypers testing different language, speech and vision providers
  • Engineers deploying interactive media agents with server and Kubernetes tooling

Documented uses

  • Combine speech recognition, language models and speech synthesis in a live agent.
  • Process video with YOLO, Roboflow or custom PyTorch and ONNX models.
  • Add retrieval, memory, tool calling or MCP actions to an agent conversation.
  • Connect agents to phone calls through Twilio or Telnyx.
  • Serve and monitor agent workloads with HTTP, Prometheus and scaling features.
Before you choose

Requirements & limits

The repository is public and licensed under Apache-2.0. The documented Stream-based setup requires Stream API credentials. Installation is shown through uv, and the framework is intended for Linux, Windows, macOS, CLI and server or self-hosted environments.

  • Provider flexibility: the framework lists integrations across LLM, realtime, speech-to-text, text-to-speech and video-processing services.
  • Video pipeline control: processors can run YOLO, Roboflow or custom PyTorch and ONNX models before or after model calls.
  • Conversation tooling: VAD, diarization, turn detection, tool calling, MCP and text back-channel messaging are listed features.
  • Deployment scope: built-in HTTP serving, Prometheus metrics, horizontal scaling and Kubernetes deployment support production-oriented workflows.
  • Media transport: WebRTC is used for real-time video, while Stream’s edge network is documented as one low-latency option.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
2 October 2026
Original website title
GetStream/Vision-Agents · Official homepage title · retained in its source language
How we review sourcesSuggest a correction