multimodal-agent 开源项目推荐（38 个）

X-PLUG /MobileAgent

移动代理：强大的图形用户界面代理家族

om-ai-lab /OmAgent

[EMNLP-2024] Build multimodal language agents for fast prototype and production

bz-lab /AUITestAgent

AUITestAgent 是首个自动化、基于自然语言驱动的移动应用 GUI 测试工具，能够完全自动化整个 GUI 交互和功能验证过程。

MadeAgents /mobile-use

MobileUse：一款开源的移动图形界面代理，用于Android手机自动化、AndroidWorld/AndroidLab评估、层级反思和主动探索。

zjunlp /LightMem-Ego

[EMNLP 2026] LightMem-Ego: Your AI Memory for Everyday Life

kwai /MobileForge

Official code for "MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization"

buptlg /fireagent

Knowledge-enhanced multimodal Fire Safety Agent with two-hop reasoning, evidence-tree RAG, open-set guardrails, and four-layer memory.

syrin-labs /syrin-harness

The Python Harness for Production AI Multi-Agent Systems

bilgeyucel /multimodal-agent-workshop

🖼️ Workshop: Build a multimodal AI agent with Haystack & GPT-4o — featuring image understanding, document retrieval, conversational memory, and human-in-the-loop safety controls

jiangjin1999 /awesome-multimodal-agentic-coding

Papers, benchmarks, demos, and tools for multimodal coding agents: visual feedback for UI, CAD, 3D, games, charts, and robotics.

Raidriar7170 /browser-use-vision

Vision-grounding plugin for browser-use agents with SoM, Florence-2, Vision–DOM alignment, adaptive visual context, and objective evaluation.

bigai-nlco /ExoViP

[COLM 2024] ExoViP: Step-by-step Verification and Exploration with Exoskeleton Modules for Compositional Visual Reasoning

ambient-intelligence-hq /ambient

A video understanding and research agent that can reason over long-form videos

UCSB-NLP-Chang /WorldAuditBench

Official Repository of WorldAuditorBench

dikshant182004 /MathTutor

🧮 Multi-agent AI math tutor built with LangGraph — CRAG retrieval, episodic & semantic long-term memory, Tavily MCP web search, Google OAuth, and Neo4j-style memory graph. Powered by LLaMA 3.3 70B on Groq.

Kshitijm7 /digital-persona

A persistent, emotionally reactive 3D Digital Persona powered by Gemini 2.5 Flash Native Audio. Features sub-100ms conversational latency and procedural ARKit emotive realism.

MrGood4U /cosight

Cosight — A realtime multimodal AI agent runtime with composable roles and capabilities.

KID0365 /PaperPilot

LangGraph-based multimodal paper reading agent for AI/ML papers, with PDF parsing, visual extraction, memory, and Markdown/PDF report generation.

MathAdventurer /video-security-agent

Evidence-grounded video security agent powered by Qwen3-VL, FastAPI, adaptive sampling, policy guards, and auditable sessions

AbhyudayPatel /DuoTalk

Multi-Agent voice conversation platform