Rome OS: Compounding Agent OS
Rome OS is a compounding agent operating system designed for recursive agents. It positions itself as an open-source alternative to platforms like Grok Bot and Meta's Muse, built with TypeScript.
Hippo Memory v1.57.0: AI Agent Memory with Self-Correction
Hippo Memory v1.57.0 introduces an AI agent memory system that learns from errors, preventing repetition. It utilizes a local SQLite store and MCP server for persistence and integrates with various AI coding tools.
BrowserOS: Open-Source Agentic Browser
BrowserOS is an open-source agentic browser designed as an alternative to platforms like ChatGPT Atlas and Perplexity Comet. It supports multiple operating systems and integrates with various LLM providers.
DSPy Framework for TypeScript - Latest Release
The ax-llm/ax repository provides the DSPy framework for TypeScript, with its latest release, version 24.0.24, published on September 27, 2026. This framework is designed for building AI applications using large language models.
Motita: Pure Go Autonomous CLI Agent with Deterministic Validation
Motita is a new autonomous CLI agent built entirely in Go, featuring a unique three-layer architecture that separates task proposal, execution, and validation. This design ensures that tasks are only declared complete after a deterministic anchor, implemented by the user's code, verifies the outcome.
new-acad: AI Agent for AutoCAD/Civil 3D
The new-acad project introduces an in-app AI agent for AutoCAD and Civil 3D, integrating LLM capabilities directly into the CAD environment. It offers a palette for natural language interaction and provides access to over 270 MCP tools and 390+ plugin methods for complex engineering tasks.
Servosity Releases Open-Source MSP AI Skills Toolkit
Servosity has released msp-skills, an open-source toolkit enabling AI models like Claude and ChatGPT to interact with Managed Service Provider (MSP) tools. The toolkit offers local-first data processing and includes connectors for PSA, RMM, backup, and M365 applications.
Buddy Android: Floating Chat Bubble for AI Phone Control
Buddy Android is a floating chat bubble application that enables AI models like Claude Code, Codex, or OpenCode to interact with and control an Android phone. It operates directly within the app, supporting both paid and free AI models without requiring ADB, a PC, or root access.
Happy: Orchestrate Claude, Codex, Grok
Happy is an open-source application designed to orchestrate AI models like Claude, Codex, and Grok. It allows users to leverage existing subscriptions for these services. The project recently released version cli-1.2.6-beta.1.
Sibyl v1.4.10: Cross-Agent Memory for AI Coding Tools
Sibyl, a self-hostable knowledge graph, has released version 1.4.10. It provides cross-agent memory for AI coding tools like Claude Code, Codex, and Cursor, enabling shared knowledge across different agents.
Vetto v0.6.3: Daemon-less Sandbox for AI Coding Agents
Vetto, a daemon-less OS security boundary and sandbox for AI coding agents, has released version v0.6.3. Written in Rust, it aims to provide a secure environment for tools like Codex, Claude Code, Cursor, and Aider.
AI DevKit: Control Plane for AI Coding Agents
AI DevKit is a TypeScript project serving as a control plane for AI coding agents. It recently released version cli@0.69.0, indicating active development and a focus on developer tools.
Agent Teams AI: Multi-Agent Orchestration with Kanban Board
Agent Teams AI is a TypeScript project enabling multi-agent systems where agents communicate and review each other's work. It offers a Kanban board interface for high-level command and oversight, supporting over 300 models and 200 LLM providers.
HealthClawGuardrails: AI Agent Security for FHIR Data
HealthClawGuardrails is an open-source project providing security layers for AI agents interacting with FHIR clinical data. It offers features like PHI redaction, immutable auditing, and tenant isolation, with recent updates including demo videos.
yoagent v0.24.0: Rust Agent Loop with 7 LLM Protocols
The yoagent Rust library, version v0.24.0, provides an agent loop capable of streaming from seven LLM protocols, executing tools, and continuing until a task is completed. This project has seen a recent release and has accumulated 280 stars and 40 forks.
Ruby LLM v2.0.0: Ruby-Native AI Framework
The Ruby LLM framework, now at v2.0.0, offers a Ruby-native approach to AI capabilities including chats, agents, and multimodal processing. It aims to provide a consistent API for various AI models within Ruby or Rails applications.
Blave Agent v0.1.11: AI Quant Infrastructure for macOS
Blave Agent has released desktop-v0.1.11, an open-source macOS workspace for AI agents focused on quantitative trading infrastructure. It enables AI models like Claude Code or Codex to transform trading ideas into backtested strategies and execute them live.
claude-mem: Persistent Context for AI Agents
The claude-mem project provides persistent context across sessions for AI agents. It captures agent interactions, compresses them using AI, and reintroduces relevant context into future sessions. This system supports a range of AI models including Claude Code, Gemini, and Copilot.
CareerOps AI Job Search Agent
CareerOps is an open-source AI job search agent designed to run locally within your AI coding CLI. It scans job portals, evaluates listings, tailors your CV, and tracks applications, with its latest release being web-v0.12.0.
OpenSider for VS Code Integrates Multiple AI Agents
OpenSider for VS Code is a new open-source extension that allows developers to use multiple AI agent CLIs, such as Claude Code, Codex, Cursor, OpenCode, and GitHub Copilot CLI, from a single VS Code side panel. It offers a unified interface for these agents without bundling any models itself.
Jarvis AI Agent: Self-Hosted Linux Automation with Multi-LLM Support
Jarvis is a self-hosted, autonomous AI agent for Linux that can control the desktop via VNC, integrate with WhatsApp, and utilize a RAG knowledge base. It supports multiple LLMs, including local Ollama models, and features a sandboxed security layer for multi-user environments.
Felix: Self-Hostable Managed Agents Harness
Felix is a self-hostable agents harness that allows users to author agents via YAML manifests. These manifests are compiled into governed agents with features like durable fibers, memory, skills, evaluation, approvals, and sandboxes, supporting multiple LLM backends. The system is designed for flexible deployment across Docker, Helm, AWS, or GCP.
Wippy AI Releases Bee: Distributed Workspace for AI Agents
Wippy AI has released Bee, a distributed terminal desktop designed to manage coding agents and applications within a unified workspace. Bee operates on the Wippy runtime, allowing processes like shells and agents to run independently and persist even after a terminal session closes.
PI-Desktop: Local-First AI Coding Agent Workspace
PI-Desktop is a new, local-first desktop application designed to provide AI coding agents with a persistent and extensible workspace. It supports model-agnostic workflows, user-installable plugins, and advanced agent orchestration capabilities.