HomeProjectsArticlesPK BoardAbout
Explore Projects

Vision Agents

Active
GitHub Python Apache-2.0

Description

Open Vision Agents by Stream. Build voice and vision agents quickly with any model or video provider, using Stream's edge network for ultra-low latency realtime interactions.

Key Features

  • Real-time video AI with YOLO, Roboflow integration combined with Gemini/OpenAI for live visual understanding
  • Ultra-low latency audio/video under 30ms via Stream's edge network with 500ms join time
  • Native SDK methods from OpenAI, Gemini, and Claude for always-latest LLM capabilities
  • Pluggable video processor pipeline supporting custom PyTorch/ONNX models alongside LLM calls
  • Multi-platform SDKs for React, Android, iOS, Flutter, React Native, and Unity
  • Production-ready with HTTP server, Prometheus metrics, horizontal scaling, and Kubernetes deployment

Use Cases

💡 Real-time video coaching applications (sports, physical therapy, workout guidance)
💡 Multi-modal AI assistants that see, listen, and respond to video streams
💡 Drone fire detection and surveillance with live visual AI processing
💡 Phone-based voice agents with video understanding via Twilio integration

Strengths & Limitations

✅ Strengths

  • • Actively maintained, recent updates
  • • High community interest (8.1k stars)
  • • Permissive open-source license (Apache-2.0)
  • • Responsive to issues, low backlog

Tags

voice agent python llm framework web

Categories

🌐 Browser Agent ⚡ Agent Tools

Quick Start

Install: uv add vision-agents. Get a Stream API key. Follow the quickstart guide at visionagents.ai to build your first voice or video agent.
Visit GitHub

Related Projects

MiroThinker

8.4k · Python
Normal A

A deep research agent framework optimized for complex research and prediction tasks, with MiroThinker-1.7 and MiroThinker-H1 models achieving 74.0 and 88.2 on BrowseComp benchmark, supporting multi-step reasoning and information retrieval.

pythonagentllm +3
  • · Deep research agent framework optimized for complex research and prediction tasks
  • · MiroThinker-1.7 achieves 74.0 on BrowseComp, H1 model reaches 88.2
  • · Supports 256K context window and up to 300 tool calls per task
View full details → View on GitHub

Tongyi DeepResearch

19.9k · Python
Stale B

Open-source deep research agent from Alibaba Tongyi Lab, using multi-stage iterative information retrieval and reasoning to conduct deep analysis, synthesis, and summarization of complex topics with web search and document analysis.

pythonagentllm +3
  • · 30.5B total parameter agentic LLM with only 3.3B activated per token (MoE architecture)
  • · Fully automated synthetic data generation pipeline for agentic pre-training, SFT, and RL
  • · Large-scale continual pre-training on diverse agentic interaction data
View full details → View on GitHub