AIエージェント
LocalAI
LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
★ 49.4k スターMIT
ガイドを見る →Vision-Agentsは、GetStreamが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: GetStream/Vision-Agents）。GitHubスターは8,154、フォークは692で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2026年10月8日です。
Stream's open Vision Agents — build real-time voice and vision agents with any model or video provider.
Pick your video provider, the models you will combine (for example a detector plus an LLM), and get Stream credentials if you use Stream's edge network.
Clone GetStream/Vision-Agents from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Add vision-agents with uv, optionally with the extras bundle for Stream, OpenAI, ElevenLabs and Deepgram integrations.
Follow the quickstart to join a call with a basic agent — the repo's golf-coach example shows the pattern — and have it respond to what it sees and hears.
Built for live video: combine YOLO, Roboflow and others with Gemini or OpenAI in real time.
Uses Stream's edge network with fast joins and very low audio/video latency, while remaining open to other video providers.
Work through native SDK methods from OpenAI, Gemini and Claude rather than a lowest-common-denominator wrapper.
SDKs for React, Android, iOS, Flutter, React Native and Unity powered by the same realtime stack.
Speech-to-text, text-to-speech and vision integrations (extras include ElevenLabs and Deepgram) for agents that listen and speak as well as watch.
Example agents, including a golf coach, show full patterns for perception-plus-LLM video agents.
Get the source from the official GitHub repository for Vision-Agents.
git clone https://github.com/GetStream/Vision-Agents.git
cd Vision-AgentsInstall the core package as the README's getting-started step 1 shows.
uv add vision-agentsInstall the extras bundle for the providers you will use, then get a Stream API key and follow the quickstart to build your first agent.
uv add "vision-agents[getstream, openai, elevenlabs, deepgram]" # then add your Stream API credentials — see READMEGetStream (Stream), through the GetStream/Vision-Agents repository, with the site at visionagents.ai.
Yes — the project is open source under the Apache-2.0 license as recorded for this listing. See the license terms for any conditions on redistribution.
You need credentials for the services you use — for Stream, the README notes a free API key with monthly participant minutes; model providers such as OpenAI or Gemini use your own keys.
Upload-and-analyse APIs are offline and high-latency. Vision Agents is built for live sessions — streaming video, sub-second joins and realtime responses — which suits interactive agents but means running a realtime stack, not a batch job.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — follow the quickstart to join a call with a basic agent — the repo's golf-coach example shows the pattern — and have it respond to what it sees and hears.
LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
Real-time voice runtime that keeps AI agents talking, working, and present.
Tutti — a real-time shared workspace where people and multiple AI agents build in tune, sharing context, files, apps and tasks
Real-time knowledge graphs that give agents a memory with structure.
An open-source SDK for production agents on any model, any cloud.
Build voice-based LLM agents with a modular, open-source toolkit.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — Vision-Agentsが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でVision-Agentsとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
NVIDIA Isaac GR00T N1.7 — an open foundation model for generalist robots, ready to fine-tune and deploy.
All your agents and workspaces in one place — on every device you own.
Isaac Lab — NVIDIA's unified framework for robot learning, with 30+ environments and multi-physics simulation.
Osaurus — own your AI: a native macOS agent harness with any model, memory and fully offline execution.
FIRERPA / LAMDA — full-stack Android device control: remote desktop, automation, MITM, Frida and MCP.
Production-ready agentic RAG behind a clean REST API.