ByteDance

An acceleration library that supports arbitrary bit-width combinatorial quantization operations

Official implementation of ATI: Any Trajectory Instruction for Controllable Video Generation. https://arxiv.org/pdf/2505.22944

通用的Android inline hook库，支持thumb，arm32，arm64

Appshark is a static taint analysis platform to scan vulnerabilities in an Android app.

Bernini is a unified framework for video generation and editing that combines an MLLM-based semantic planner with a DiT-based renderer.

🔥🔥 btrace (AKA RheaTrace) is a high-performance Android & iOS tracing tool built on Perfetto. It not only times your methods but also reveals why they’re slow.

A model compilation solution for various hardware

optimized BERT transformer inference on NVIDIA GPU. https://arxiv.org/abs/2210.03052

Cloud Shuffle Service(CSS) is a general purpose remote shuffle solution for compute engines, including Spark/Flink/MapReduce.

[CVPR 2025 Highlight🌟] Official ComfyUI implementation of "HyperLoRA: Parameter-Efficient Adaptive Generation for Portrait Synthesis"

🔥 [ICCV 2025 Highlight] Official ComfyUI native node supporting InfiniteYou with FLUX

A LangChain-based framework for building super agents.

The official repo for “Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting”, ACL, 2025.

Fastbot(2.0) is a model-based testing tool for modeling GUI transitions to discover app stability problems

[English](https://github.com/bytedance/flow-builder/blob/main/README.md) | 简体中文

KV cache store for distributed LLM inference

🔥 [ICCV 2025 Highlight] InfiniteYou: Flexible Photo Recrafting While Preserving Your Identity

Inter-process message bus

JAX accelerated Quantum Monte Carlo

A 3B-active-parameter native unified multimodal model for image and video understanding, generation, and editing.

Taming Stable Diffusion for Lip Sync!

Implementation of paper: Flux Already Knows – Activating Subject-Driven Image Generation without Training

Libtpa(Transport Protocol Acceleration), a DPDK based userspace TCP stack implementation.

Lynx: Towards High-Fidelity Personalized Video Generation

MoMA: Multimodal LLM Adapter for Fast Personalized Image Generation

A simple and easy-to-use Go mocking library derived from ByteDance's internal best practices

This repo contains the code for our paper Towards Open-Ended Visual Recognition with Large Language Model

Structural Quality Assessment for Biomolecular Structure Prediction Models

PaSa -- an advanced paper search agent powered by large language models. It can autonomously make a series of decisions, including invoking search tools, reading papers, and selecting relevant references, to ultimately obtain comprehensive and accurate results for complex…

PatchEval: A New Benchmark for Evaluating LLMs on Patching Real-World Vulnerabilities

Official Repo For Pixel-LLM Codebase: Sa2VA (T-PAMI-26), SAMTok (CVPR-26), VRT (Arxiv-25), SaSaSa2VA (1-st solution for LSVOS)

A new multi-shot video understanding benchmark Shot2Story with comprehensive video summaries and detailed shot-level captions.

[ICCV 2025] Code & Data for: SuperEdit - Rectifying and Facilitating Supervision for Instruction-Based Image Editing

Tarsier -- a family of large-scale video-language models, which is designed to generate high-quality video descriptions , together with good capability of general video understanding.

A RocksDB compatible KV storage engine with better performance

The official code for NeurIPS 2024 paper: Harmonizing Visual Text Comprehension and Generation

Trae Agent is an LLM-based agent for general purpose software engineering tasks.

The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra

Pioneering Automated GUI Interaction with Native Agents

[CVPR 2026] 🔥🔥 Official Repo of UMO: Scaling Multi-Identity Consistency for Image Customization via Matching Reward

[ICCV 2025] 🔥🔥 UNO: A Universal Customization Method for Both Single and Multi-Subject Conditioning

[ICLR 2026] Official repo for paper "Video-As-Prompt: Unified Semantic Control for Video Generation"

videx, the Disaggregated, Extensible Virtual Index Engine for What-If Analysis

Web-Bench is a benchmark designed to evaluate the performance of LLMs in actual Web development.

[CVPR 2025 Highlight] X-Dyna: Expressive Dynamic Human Image Animation

[NeurIPS 2025] Official implementation of "XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation".

A HTML5 video player with a parser that saves traffic

HTML5 danmu (danmaku) plugin for any DOM element

FlowGram is an extensible workflow development framework with built-in canvas, form, variable, and materials that helps developers build AI workflow platforms faster and simpler.

Enterprise-oriented Generic Proxy Solutions. Maintenance mode now, please switch to use [VEY](https://github.com/VEY-OSS/vey) for new development efforts

Cloud-native container hardening for Kubernetes — from syscall to protocol, from workload to AI Agent.

video-SALMONN 2 is a powerful audio-visual large language model (LLM) that generates high-quality audio-visual video captions, which is developed by the Department of Electronic Engineering at Tsinghua University and ByteDance.