Open Source Project

An AI-powered security testing tool for authorized targets that combines proxy, browser, and code inspection capabilities, offering both open-source deployment and cloud services.

A speech recognition API for Swift developers that provides unified access to Apple, Parakeet, and Qwen3-ASR, and supports speaker separation.

Skyhook has launched a Kubernetes visualization and AI-powered operations tool that consolidates resource topology, event timelines, and MCP investigations.

The MiniMax team has released an open-source programming agent workflow and self-assessment toolkit that covers review, context, and long-running tasks.

An open-source document analysis and quality assessment model that integrates scanned documents, tables, and formulas into a verifiable data workflow.

Provide reusable project context, specialized workflows, and review stages for existing AI assistants, enabling collaboration across sessions and tools.

An open-source framework for optimizing language model programs using modules, data, and evaluation metrics

An Open-Source Python Orchestration Framework for Retrieval-Augmented Generation and Agents

An open-source platform for visually building AI agents, RAG, and MCP workflows

Baidu has open-sourced a document analysis model that supports multi-page joint recognition, cross-page tables, and the processing of extremely long documents.

An open-source AI-powered research and report generation agent that automatically retrieves information from the web and local sources, compiles research reports with citations, and supports self-deployment and integration with Python.

An Apache 2.0 open-source platform for document understanding, retrieval, and enhanced generation, supporting complex data parsing, knowledge base retrieval, and agent workflows.

An open-source multi-model chat and agent platform from MIT that centrally manages models, document-based Q&A, and tool invocations, and supports self-hosting by teams.

Formerly known as Perplexica, this self-hosted, open-source AI search engine combines online search with local or cloud-based models to generate answers with citations.

Parses complex documents—such as PDFs, Office files, and scanned documents—into structured data, supporting both on-premises deployment and RAG workflows.

An open-source web scraping tool designed for large language models and agents that converts authorized web pages into Markdown and structured data.

An open-source runtime for large-scale concurrent agents that allows tasks that are waiting to resume execution based on events.

HeyGen is an open-source agent-based video creation tool that uses HTML, styles, and scripts to organize scenes and render videos.

The open-source AI assistant launched by Nous Research supports skill acquisition, cross-session memory, and multi-platform task access.

Tencent's open-source knowledge platform integrates document search, agent tasks, and Wiki organization into a single knowledge base.

An open-weight model for locally programming agents, designed to work with Pi, OpenCode, and OpenHands for code understanding, modification, and testing.

Run open models on-premises or in the cloud, with desktop chat, command-line, and development interfaces available, making it ideal for local AI experiments and application integration.

A "local-first" shared memory app launched by Tencent Hunyuan allows authorized AI tools to access conversation and coding context; it is currently still in preview.

OpenCode is an open-source AI programming agent launched by Anomaly that assists users in understanding projects, modifying code, and reviewing changes via terminal, desktop, and web interfaces, and supports integration with multiple models. This article covers how to use V2, pricing limits, and recent updates from October 2026.

Paperclip centralizes the management of tasks, roles, budgets, approvals, and operational logs for multiple AI agents. It supports self-hosting and is ideal for individual developers and small teams to collaborate on AI projects.

NVIDIA OpenShell provides a controlled execution environment for AI agents through sandboxing, file and network policies, and credential proxy, making it ideal for development teams to deploy and manage permissions.

The main agent breaks down tasks, coordinates research, writing, administrative, and development assistants, and handles multi-step workflows continuously within the desktop.

An open-source, AI-driven PowerPoint generation workflow that automatically converts various types of documents into natively editable PowerPoint files, rather than static images.

A multi-agent collaboration framework jointly open-sourced by Ant Group and Renmin University of China, it schedules search agents much like an operating system and efficiently completes complex, long-duration deep information retrieval tasks through global shared state and pipelined parallel mechanisms.

Ant Group has launched a new-generation spatial perception model for robots that enhances 3D depth perception capabilities and improves performance in grasping, navigation, and environmental understanding.