* [Cherry Studio v1.8.0](https://github.com/CherryHQ/cherry-studio) – Desktop client supporting multiple LLM providers, available on Windows, Mac, and Linux. * [MCP Toolbox for Databases v0.29.0](https://github.com/googleapis/genai-toolbox) – Toolbox simplifies database interactions for AI tools, enhancing development speed and security. * [PUAX 2.0 - AI Agent v2.0.0](https://github.com/linkerlin/PUAX) – PUA-prompt toolkit for guiding AI agents using trigger detection, role recommendation, and context-based activation. * [Giselle v0.73.0](https://github.com/giselles-ai/giselle) – Platform for building and composing agentic AI workflows with a visual drag-and-drop agent builder. * [debug-that v0.5.0](https://github.com/theodo-group/debug-that) – Debugger CLI for AI agents offering token-efficient state inspection, breakpoints, and runtime-adaptive debugging using short reference handles. * [PageAgent v1.5.8](https://github.com/alibaba/page-agent) – Client-side JavaScript GUI enabling natural-language control and DOM interaction on webpages. * [BitFun v0.2.0](https://github.com/GCWing/BitFun) – Next-generation agent system providing personal AI assistants with long-term memory, personality, and built-in coding and knowledge-work capabilities. * [MoLing v0.5.0](https://github.com/gojue/moling) – Dependency-free local MCP server enabling OS-level file operations, command execution, and browser control for office automation. * [OpenPencil v0.4.0](https://github.com/ZSeven-W/openpencil) – AI-native design tool that turns prompts into live UI on a collaborative canvas with multi-agent orchestration and code export. * [OpenClacky v0.9.0](https://github.com/clacky-ai/openclacky) – Creator-focused platform for packaging encrypted, white-labeled AI workflows into sellable Skills with built-in storefront and cross-platform distribution. * [AIDD Framework v2.8.0](https://github.com/paralleldrive/aidd) – Standardized architecture, workflows, and tools for building secure, maintainable AI-driven development pipelines. * [Agentica, AI Function Calling Framework v0.44.0](https://github.com/wrtnlabs/agentica) – Agentic AI framework specialized in AI Function Calling using TypeScript classes, OpenAPI, and MCP protocols. * [@keyid/agent-kit v0.2.0](https://github.com/KeyID-AI/agent-kit) – MCP-based email toolset offering send, receive, reply, search, contacts, auto-reply, scheduling, and webhooks for AI agents. * [Acontext ui/v0.1.11](https://github.com/memodb-io/Acontext) – Context data platform for storing contexts, observing tasks, and enabling agent self-learning through experience collection. * [Inkeep Agents v0.58.19](https://github.com/inkeep/agents) – Platform offering a no-code visual builder and TypeScript SDK for creating and managing AI agents with full two-way sync. * [Agentrove v0.1.12](https://github.com/Mng-dev-ai/agentrove) – Self-hosted Claude Code workspace with multi-provider routing, sandboxed execution, and integrated web IDE. * [Parchi v0.5.1](https://github.com/0xSero/parchi) – AI-powered browser copilot for chat-driven sidebar automation to navigate, read, click, and extract via natural language. * [Continue v1.2.17-vscode](https://github.com/continuedev/continue) – Enables developers to create, share, and use custom AI code assistants with IDE extensions and a hub of models, rules, and tools. * [UnicomAI/wanwu v0.4.3](https://github.com/UnicomAI/wanwu) – Enterprise-grade multi-tenant platform for developing large model applications including intelligent agents, workflows, and text Q&A with model management. * [ComputeSDK @computesdk/workbenc...](https://github.com/computesdk/computesdk) – Provider-agnostic TypeScript toolkit for securely running untrusted code in isolated sandboxes. * [Jean v0.1.30](https://github.com/coollabsio/jean) – Desktop AI assistant for managing multiple projects, git worktrees, chat sessions, terminals, and GitHub integrations via Claude CLI. * [layrr v0.26.3](https://github.com/thetronjohnson/layrr) – CLI that injects a browser overlay to select page elements and apply natural-language edits directly to source code. * [Eliza v2.0.0-alpha.13](https://github.com/elizaOS/eliza) – Framework for developing and deploying multi-agent systems with real-time communication and extensible integrations. * [AgentField v0.1.59](https://github.com/Agent-Field/agentfield) – Control-plane infrastructure that deploys, scales, and secures autonomous AI agents as observable, identity-aware backend services. * [Cloudflare Sandbox SDK @cloudflare/sandbox@...](https://github.com/cloudflare/sandbox-sdk) – Run untrusted code in isolated, secure containers at the edge with command execution, file access, and service previews. * [Aurogen v0.2.0post2](https://github.com/UniRound-Tec/Aurogen) – Complete modular reimplementation of the OpenClaw paradigm enabling parallel multi-agent deployments with dynamic web-based configuration and ecosystem compatibility. * [Cyrus v0.2.34](https://github.com/ceedaragents/cyrus) – Claude Code-powered Linear agent that monitors assigned issues, runs isolated Git worktrees, processes them locally, and posts comments back to Linear. * [OpenClacky v0.8.6](https://github.com/clacky-ai/open-clacky) – CLI for scaffolding production-ready Rails full-stack web apps with one-click deployment, isolated environments, and AI-focused workflows. * [WuKongIM v2.2.4-20260313](https://github.com/WuKongIM/WuKongIM) – High-performance universal communication service supporting instant messaging, notifications, IoT, audio/video signaling, live streaming comments, customer service, and distributed messaging. * [Cossistant @cossistant/core@0.1...](https://github.com/cossistantcom/cossistant) – Chat support widget for React with headless components, real-time messaging, and backend infrastructure. * [Voquill desktop-dev-v0.0.494](https://github.com/josiahsrc/voquill) – Cross-platform speech-to-text workspace for dictating into any desktop app, AI cleanup of transcripts, and synchronized personal glossary. * [Tingly Box v0.260315.1200-previ...](https://github.com/tingly-dev/tingly-box) – High-performance desktop LLM proxy that unifies access to many models and providers via a single OpenAI-compatible API.