InsForge is an open-source backend platform for AI programming agents. Through MCP Server and CLI interfaces, the tool allows agents such as Cursor, Claude Code, and Copilot to autonomously manipulate data, much like backend engineers...
AudioLib is an audio version of OpenRouter launched by Yang Yue and the 43Music team. It provides developers with access to over 100,000 original music tracks created by the team over a year through an API. No SDK, player, copyright negotiations, or music library are required...
AI Job Search is an open-source AI job search framework based on Claude Code. It uses `/setup` to create detailed professional profiles, `/scrape` to intelligently assess job matching, and `/apply` to initiate a 'draft-review' dual-agent workflow, automatically generating...
HiDream-O1-Image is an open-source 8-bit pixel-level native unified image generation model from Zhixiang Future. It employs a world-first UiT architecture, enabling direct generation within a single token space without the need for a VAE or a separate text encoder...
OpenAI CLI is an official command-line interface tool released by OpenAI, spearheaded by Jason Liu, a developer experience engineer from the Codex team.
Camofox Browser is an open-source AI Agent anti-detection browser server built on Camoufox (a heavily modified C++ version of Firefox with fingerprinting capabilities). It provides AI with structured webpage snapshots (reduced in size) via a REST API...
Multica is an open-source AI agent team collaboration platform founded by Jiayuan, a former TikTok engineer. The platform transforms coding agents like Claude Code, Codex, and Kimi into real teammates, enabling them to collaborate on Kanban boards...
WebWorld is a series of large-scale web world models open-sourced by Alibaba's Qwen team. Trained on the Qwen3 platform, it includes 8B, 14B, and 32B versions. The model predicts web page states within a simulated browser environment...
TACO is an open-source, plug-and-play, end-agent self-evolutionary observation compression framework that requires no training and is designed for use by the University of Manchester, Beijing University of Aeronautics and Astronautics, Hong Kong University of Science and Technology, and the MAP team.
Mirage is a unified virtual file system for AI agents developed by the strukto-ai team. The tool mounts heterogeneous data sources such as S3, Slack, Gmail, GitHub, and MongoDB into a single virtual file tree...
SkillClaw is an open-source AI Agent skills collective evolution framework developed by AMAP-ML (Gaode Maps Machine Learning Team). The framework intercepts Agent sessions through a local API proxy and automatically extracts and processes skills from real interactions in the background...
GPT-5.5-Cyber is a cybersecurity-specific model from OpenAI, offering a limited preview for authenticated defenders. Built on GPT-5.5, it reduces security limitations through a trusted network access framework and supports vulnerability...
StepAudio 2.5 Realtime is an end-to-end real-time voice model developed by StepAudio, focusing on a human-like voice dialogue experience. The model supports deep content-level interaction and its voice performance is completely close to that of a real person, possessing top-tier capabilities...
Ring-2.6-1T is a trillion-parameter deep inference model launched by Ant Financial. Belonging to the Ring series, it focuses on slow-thinking capabilities for complex cognitive tasks such as math competitions and code generation. The model employs a MoE hybrid expert architecture, excelling in high-parameter...
ERNIE 5.1 is Baidu's flagship large language model, based on ERNIE 5.0 knowledge distillation and the Once-for-All elastic training framework, reducing the pre-training cost to 6% of the industry average for similar models.
Doubao-Seed-2.0-lite is the first full-modal understanding model launched by ByteDance's Doubao team. The model supports native unified understanding of video, images, audio, and text, and simultaneously upgrades the Agent, Coding, and GUI capabilities.
GPT-Realtime-Translate is a real-time speech translation model launched by OpenAI. It supports real-time translation of over 70 input languages into 13 output languages, employing an end-to-end architecture to directly process the raw audio, skipping the intermediate text processing...
GPT-Realtime-Whisper is a streaming speech-to-text model launched by OpenAI. It is an upgrade based on the Whisper architecture and can output text in real time while the user speaks, with extremely low latency.
GPT-Realtime-2 is OpenAI's second-generation real-time speech model and currently the most powerful speech agent in the Realtime API. The model boasts GPT-5 level inference capabilities, enabling it to listen and reason simultaneously during speech conversations...
SenseNova 6.7 Flash-Lite is a new generation of lightweight multimodal intelligent agent model launched by SenseTime. It adopts a native multimodal architecture and is designed for real-world workflows.
Grok Voice Think Fast 1.0 is xAI's flagship voice agent model, specializing in complex, multi-step real-world scenarios. The model ranks first in the τ-voice Bench, supports 25 languages and full-duplex dialogue, and features low latency...
Realtime TTS-2 is a next-generation real-time speech synthesis model from Inworld AI, specifically designed for conversational AI scenarios. The model can convert text into natural speech and, more importantly, 'understand' the audio emotion, tone, and... of the conversational context.
GenericAgent (GA) is a general-purpose, self-evolving LLM agent system developed by A3 Lab, with 'maximizing contextual information density' as its core design principle. GenericAgent utilizes 9 atomic tools and a 30k context window...
Models.dev is an open-source AI model database from the OpenCode team. Models.dev integrates model information from dozens of major vendors, including OpenAI, Anthropic, and Google, in a standardized TOML format, covering aspects such as price, ...