audio-understanding 开源项目推荐（18 个）

HKUDS /VideoAgent

VideoAgent：用于视频理解、编辑和重新制作的一体化学习框架

OpenMOSS /MOSS-Audio

MOSS-音频是一个开源基础模型，用于统一音频理解，支持现实世界场景中的语音、声音、音乐、字幕、QA和推理。

oxbshw /watch-skill

Give AI agents eyes, ears, and verifiable results. Watch Skill turns video, audio and screen activity into searchable, timestamped evidence and proves work with deterministic contracts, not model opinion. DeepWatch is the agent workspace built on DeepSeek Harness. Python + npm, MCP, CLI, REST, Web.

TaoLiveAIGC /TLive-Omni

TLive-Omni: An Omni-Modal Understanding Model for E-Commerce Live Streaming

NKU-HLT /DIFFA

[AAAI 2026 & ACL 2026] The official implementation of the DIFFA series for dLLM-based large audio language model

JHU-LCAP /FlexSED

open-vocabulary sound event detection

khalooei /Voxtral-AI-Demo-Local-Interface

Voxtral is a state-of-the-art model developed to handle both speech transcription and audio understanding with remarkable accuracy and efficiency. This demo interface lets you run the Voxtral model on powerful GPUs to evaluate its performance and see how it can be used for transcription and deeper analysis.

OpenMOSS /FutureOmni

用于评估多模态语言模型基于音频与视频上下文进行未来事件预测的基准

ZXXZ1000 /LATRACE-AI

Open-source AI memory engine for agents: multimodal memory, temporal knowledge graphs, graph RAG, and evidence-backed long-term context.

OpenMOSS /sglang

An OpenMOSS fork of SGLang with serving support for the MOSS-TTS family and MOSS-Audio

go-agent-x /VideoXAgent

Online Video Agent Harness for Long Video Understanding.

danielrosehill /Audio-Multimodal-AI-Resources

A compilation of resources (model profiles, benchmarks, docs) for multimodal AI models with audio understanding (esp. focused on ASR and transcription use-cases)

liyixuan201211 /dsh-audio-read

让 AI Agent 听懂音频：转写、鸟种识别、语音情感、环境声与音乐分析。全部本地运行，不用大模型。DeepSeek Harness 技能。

danielrosehill /Audio-Understanding-Bitrate-Eval-0426

Empirical eval: how MP3 bitrate affects transcription accuracy across every audio-input LLM on OpenRouter (Gemini, GPT-Audio, Voxtral, MiMo). April 2026.

2xmncvcx92-dotcom /guanzhe-asr-listen

Local-first OpenClaw skill for ASR, subtitles, dialect transcription, and audio understanding with whisper.cpp and Xiaomi MiMo.

snkii /awesome-lalm

Awesome Large Audio Language Models (LALMs) - a curated index of audio-understanding foundation models: papers, venues, code, and open weights

api-evangelist /gemini

Gemini — independent third-party profile of a public API surface, by API Evangelist. Google's Gemini API provides access to state-of-the-art generative AI models for text generation, multimodal understanding, code generation, and more.

danielrosehill /Gemini-31-Lite-Audio-Understanding-Eval

One voice recording for testing with TTS/cloning