Open Source Agentic AI Tools
Sort By:
Agentic AI Tools
Browse free open source Agentic AI tools and projects for Mac and ChromeOS below. Use the toggles on the left to filter open source Agentic AI tools by OS, license, language, programming language, and project status.
-
Gemini 3 and 200+ AI Models on One PlatformBuild generative AI apps with Vertex AI. Switch between models without switching platforms.
-
Try Google Cloud Risk-Free With $300 in CreditUse your credit across every product. Compute, storage, AI, analytics. When it runs out, 20+ products stay free. You only pay when you choose to.
-
1
PentestGPT
Automated Penetration Testing Agentic Framework Powered by LLMs
PentestGPT is an AI-powered autonomous penetration testing agent designed to perform intelligent, end-to-end security assessments using large language models. Published at USENIX Security 2024, it combines advanced reasoning with an agentic workflow to automate tasks traditionally handled by human pentesters. The platform supports multiple penetration testing categories, including web security, cryptography, reversing, forensics, privilege escalation, and binary exploitation. PentestGPT runs in a Docker-first environment, providing a secure, reproducible setup with built-in tooling and session persistence. It offers real-time feedback and live walkthroughs, allowing users to observe each step of the testing process as it unfolds. Built with a modular and extensible architecture, PentestGPT supports cloud and local LLMs, making it suitable for research, education, and authorized security testing. -
2
llmfit
157 models, 30 providers, one command to find what runs on hardware
llmfit is a terminal-based utility that helps developers determine which large language models can realistically run on their local hardware by analyzing system resources and model requirements. The tool automatically detects CPU, RAM, GPU, and VRAM specifications, then ranks available models based on performance factors such as speed, quality, and memory fit. It provides both an interactive terminal user interface and a traditional CLI mode, enabling flexible workflows for different user preferences. llmfit also supports advanced configurations including multi-GPU setups, mixture-of-experts architectures, and dynamic quantization recommendations. By presenting clear performance estimates and compatibility guidance, the project reduces the trial-and-error typically involved in local LLM experimentation. Overall, llmfit serves as a practical decision assistant for developers who want to run language models efficiently on their own machines. -
3
OpenClaw-RL
Train any agents simply by 'talking'
OpenClaw-RL is an open-source reinforcement learning framework designed to train and personalize AI agents built on the OpenClaw ecosystem. The project focuses on enabling agents to improve their behavior through interactive learning rather than relying solely on static prompts or predefined skills. One of its key ideas is allowing users to train an AI agent simply by interacting with it conversationally, using natural language feedback to guide the learning process. The system incorporates reinforcement learning techniques to refine the agent’s policies for tool use, decision making, and task completion over time. It also explores approaches such as online policy distillation and hindsight feedback signals to strengthen training signals from real interactions. The framework operates asynchronously and does not require external API keys, making it easier to experiment with local agent training workflows. -
4
Antigravity Kit
AI Agent templates with Skills, Agents, and Workflows
Antigravity Kit is an AI agent capability expansion toolkit that provides structured templates, skills, and workflows designed to boost the productivity of AI assistants and autonomous agents in coding environments. It comes with an extensive library of predefined agent personas and domain-specific skills that help in performing targeted tasks such as frontend development, backend engineering, quality assurance, and more. With this kit, developers can rapidly initialize a project scaffold and gain access to specialist agents and command workflows that guide the agent through multi-step tasks and slash-command procedures. The repository follows a clear pattern for defining agents and their behaviors, making it easier to extend with new skills or tailor existing ones to your use case. Because it’s framework-agnostic at its core, the Antigravity Kit can integrate with IDEs and other development tooling for smoother AI-augmented coding. -
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
5
OpenManus
Open-source AI agent framework
OpenManus is an open-source AI agent framework designed to autonomously execute complex, multi-step tasks by combining reasoning, planning, and tool use. It enables developers to build agents that can think, act, and iterate toward goals rather than simply responding to prompts. The platform emphasizes task decomposition, allowing agents to break down objectives into smaller steps and execute them sequentially or recursively. OpenManus supports integration with external tools, APIs, and environments, making it suitable for real-world automation workflows. It is built to be flexible and extensible, enabling customization of agent behaviors, tools, and reasoning strategies. Overall, OpenManus provides a foundation for creating more capable, autonomous AI systems that can handle dynamic and goal-driven tasks. -
6
Preline UI
Preline UI is an open-source set of prebuilt UI components
Preline is an open-source UI component library designed to work alongside utility-first Tailwind CSS projects, providing a comprehensive set of prebuilt, responsive, interactive interface elements for modern web development. It includes a rich collection of components such as buttons, navigation bars, dropdowns, modals, form controls, and more that are styled using Tailwind’s utility classes and are easy to customize without writing low-level CSS. Developers can quickly assemble complex, mobile-friendly user interfaces with consistent design and behavior straight out of the box, greatly reducing the overhead of crafting common UI patterns from scratch. Preline also offers setup guidance and integration examples so teams can get started rapidly within their Tailwind projects. Because it follows Tailwind’s conventions, it integrates smoothly with other Tailwind-first tools and workflows, maintaining a lean stylesheet footprint and maximizing flexibility. -
7
The Pope Bot
Autonomous AI agent that you can configure and build
The Pope Bot is an autonomous AI agent framework that lets users configure and run an AI-powered agent that can perform tasks continuously, day in and day out, by leveraging GitHub Actions, commit history, and secure workflows. It’s designed so that every action taken by the agent is logged as a git commit, giving users complete visibility into what the agent did, why it did it, and when, which makes actions auditable and reversible. The framework treats the repository itself as the agent’s “brain,” and GitHub Actions serve as the compute layer, enabling tasks to run securely without exposing sensitive API keys to the underlying AI. The system integrates with messaging platforms like Telegram, where users can interact with the bot, trigger actions, or receive notifications, and supports scheduling and automation through patterns of request handling. -
8
Lossless Claw
LCM (Lossless Context Management) plugin for OpenClaw
Lossless Claw is an advanced context management plugin for the OpenClaw agent ecosystem that redefines how conversational memory is handled in large language model systems. Instead of relying on traditional sliding-window truncation or lossy summarization, it introduces a lossless architecture that preserves all historical messages while maintaining usable context within token limits. The system stores every interaction in a persistent database and incrementally summarizes older content into a hierarchical directed acyclic graph, allowing efficient compression without discarding information. This structure enables agents to dynamically reconstruct detailed context by expanding summaries when needed, effectively simulating perfect long-term memory. -
9
Medeo Video Generator
AI-powered video generation skill for OpenClaw
Medeo Video Generator is an AI-driven project designed to enable advanced video processing and generation capabilities within agent-based or automation systems. It provides a “skill” module that can be integrated into AI agents, allowing them to create, edit, and manipulate video content programmatically. The project focuses on bridging the gap between language-based AI systems and multimedia outputs by enabling models to produce structured video content as part of their workflows. It supports tasks such as video generation, editing, and transformation, making it useful for applications in content creation, marketing, and automated media production. The framework is designed to be modular, allowing developers to plug video capabilities into larger AI pipelines or agent systems. It emphasizes ease of integration and scalability, enabling both simple use cases and more complex multimedia workflows. -
Add Two Lines of Code. Get Full APM.Works out of the box for Rails, Django, Express, Phoenix, and more. Monitoring exceptions and performance in no time.
-
10
OpenClaw Office
OpenClaw Office is the visual monitoring and management frontend
OpenClaw Office is a visual monitoring and management interface designed for the OpenClaw multi-agent system, providing an immersive and interactive way to observe and control autonomous AI agents. It presents agent activity through a virtual office environment, where each agent is represented as an animated entity within a 2D or 3D workspace. The platform enables real-time visualization of agent states, interactions, and workflows, making complex multi-agent coordination easier to understand and debug. Users can observe communication flows between agents through visual connections, track token usage and operational costs, and analyze performance through integrated dashboards and charts. The system also includes live chat capabilities, allowing users to monitor conversations and tool calls as they occur. -
11
OpenWork AI
An open-source alternative to Claude Cowork, powered by opencode
OpenWork is a framework for building decentralized collaborative work environments powered by AI and human contributions. At its core, the project enables contributors to define tasks, workflows, and goals that can be split, shared, and recombined across distributed nodes while agents and humans cooperate to advance progress. It offers structured templates for work items, decision logic for task allocation, and consensus mechanisms that let groups verify and validate results toward shared objectives. This project also includes moderation and reputation layers so that contributor trust and quality can be assessed and integrated into future task assignments. Rather than a single monolithic workflow engine, it emphasizes openness — providing APIs and interfaces so communities can build custom dashboards, integrate specialized agents, or add bespoke evaluation criteria. -
12
Grok CLI
An open-source AI agent that brings the power of Grok
Grok CLI is a command-line interface built around the Grok AI model that brings programmatic and conversational AI capabilities directly to developer terminals. It lets you run Grok queries from your shell, scripting environment, or automation workflows without switching to a browser, enabling utility in scripting, quick data exploration, code generation, and assistant-guided tasks directly where you write code. The CLI supports streaming responses, so outputs appear in real time as the Grok model generates them, making interactions feel responsive and fluid in terminal contexts. Grok CLI is designed to integrate with existing terminal habits—aliases, pipes, editors, and tooling—so you can combine AI assistance with native command-line workflows like grep, awk, and git. It also includes authentication support, configuration management, and caching options so frequent queries are efficient. -
13
Pinchtab
High-performance browser automation bridge and orchestrator
Pinchtab is a lightweight browser automation backend built specifically for AI agents that need efficient, programmatic web control. Implemented as a small standalone HTTP server, it allows any agent or script to interact with web pages using simple API calls instead of heavyweight browser frameworks. The tool emphasizes accessibility-first snapshots that dramatically reduce token usage compared to screenshot-based approaches, making it cost-effective for large-scale automation. It launches and manages its own Chrome instance while remaining framework-agnostic, so it can be used with any language or agent system. Pinchtab also supports persistent sessions, stealth automation, and both headless and headed operation modes. The project’s goal is to provide fast, cheap, and portable browser control infrastructure for modern AI workflows. -
14
ClawRouter
Smart LLM router
ClawRouter is a flexible networking and routing framework designed to support AI-oriented distributed systems and agent ecosystems by managing how messages, requests, and responses are routed between components. It provides a programmable router abstraction that can handle complex traffic patterns, enabling dynamic message forwarding, load balancing, and custom routing logic based on content, context, or policy rules. Because distributed AI systems often involve many services, agents, and runtime components interacting with each other and with external APIs, ClawRouter helps ensure that communication paths remain clear, efficient, and adaptable as systems scale. The framework supports plugin-based extensions so developers can define custom protocols, transformation hooks, and monitoring handlers without modifying core routing logic. It also offers operational features like health checking, metrics reporting, and failure handling that make production deployments more reliable. -
15
AGI
The first distributed AGI system
AGI project is an experimental framework focused on building components and infrastructure for artificial general intelligence systems, emphasizing modularity, autonomy, and scalable intelligence pipelines. It aims to provide a foundation for creating agents that can reason, plan, and execute tasks across diverse domains by integrating multiple AI capabilities into a unified system. The project typically explores concepts such as agent orchestration, memory systems, task decomposition, and decision-making loops, enabling the development of more generalized and adaptive AI behaviors. It is designed to be extensible, allowing developers to plug in different models, tools, and data sources to enhance agent performance. The framework encourages experimentation with AGI-like architectures, making it useful for researchers and developers interested in advancing beyond narrow AI applications. -
16
AutoGen
An Open-Source Programming Framework for Agentic AI
AutoGen is an open-source programming framework for building AI agents and facilitating cooperation among multiple agents to solve tasks. AutoGen aims to provide an easy-to-use and flexible framework for accelerating development and research on agentic AI, like PyTorch for Deep Learning. It offers features such as agents that can converse with other agents, LLM and tool use support, autonomous and human-in-the-loop workflows, and multi-agent conversation patterns. AutoGen provides multi-agent conversation framework as a high-level abstraction. With this framework, one can conveniently build LLM workflows. AutoGen offers a collection of working systems spanning a wide range of applications from various domains and complexities. AutoGen supports enhanced LLM inference APIs, which can be used to improve inference performance and reduce cost. -
17
Claude Subconscious
Give Claude Code a subconscious
Claude Subconscious is an experimental plugin that enhances AI coding workflows by introducing a persistent “memory layer” for Claude Code through integration with Letta’s agent framework. It operates as a background agent that continuously observes user interactions, reads project files, and processes session transcripts to build long-term contextual memory. Unlike standard AI interactions that reset between sessions, this system accumulates knowledge over time, allowing it to recall user preferences, project structures, and recurring patterns across multiple sessions. The plugin injects relevant context and guidance back into Claude before each prompt, effectively “whispering” insights that improve continuity and decision-making. It also has access to tools such as file reading, code search, and web browsing, enabling it to perform background analysis and augment responses with deeper context. -
18
DenchClaw
Fully Managed OpenClaw Framework for all knowledge work ever
DenchClaw is a local-first AI-powered CRM and productivity platform built on top of the OpenClaw framework, designed to transform a user’s entire computer into a programmable, agent-driven workspace. Unlike traditional cloud-based CRMs or AI tools, it runs entirely on the user’s machine and exposes a web interface locally, allowing full control over data, workflows, and automation without relying on external servers. The system combines database management, browser automation, and AI reasoning into a unified interface where users can interact with their data and tools using natural language commands. It can ingest data from sources such as Google Drive, Notion, Gmail, and CRM platforms, consolidating everything into a centralized workspace for analysis and action. One of its most distinctive capabilities is its ability to use the user’s existing browser session, enabling it to log into services, scrape data, and perform actions like outreach or research as if it were the user. -
19
Obsidian Skills
Agent skills for Obsidian
Obsidian-Skills is a repository of agent skills tailored for use with Obsidian and any Claude-compatible agent that follows the standard Agent Skills specification, enabling AI assistants to better understand and interact with Obsidian content. These skills are markdown-driven specifications that teach Claude Code (or similar agents) how to perform context-aware tasks within Obsidian’s unique environment, such as interpreting different file types and workflows, automating workflows tied to notes, or enhancing agent responses with structured knowledge. By providing formal descriptions of patterns, conventions, and workflows common to Obsidian users, the skills empower AI tools to give more relevant suggestions, generate content that adheres to user conventions, or execute complex multi-step operations that respect the knowledge graph and file relationships. -
20
OpenAI Agent Skills
Skills Catalog for Codex
OpenAI Agent Skills is an open-source repository that serves as a broad catalog of agent skills designed to extend the capabilities of OpenAI Codex and other AI coding agents. It organizes reusable, task-specific workflows, instructions, scripts, and resources into modular skill folders so that an AI agent can reliably perform complex tasks without repeated custom prompting, making agent behavior more predictable and composable. Each skill is defined with clear metadata and instructions organizing how an AI assistant should complete specific tasks ranging from project management to code generation and documentation assistance. The repository supports community contributions, allowing developers to add new skills or update existing ones to keep the catalog relevant and practical for evolving use cases. -
21
Cloudflare Agents
Build and deploy AI Agents on Cloudflare
Cloudflare Agents is an open-source framework designed to help developers build, deploy, and manage AI agents that run at the network edge. It provides infrastructure for creating stateful, event-driven agents capable of real-time interaction while maintaining low latency through Cloudflare’s distributed platform. The project includes SDKs, templates, and deployment tooling that simplify the process of connecting agents to external APIs, storage systems, and workflows. Its architecture emphasizes persistent memory, enabling agents to maintain context across sessions and interactions. Developers can orchestrate complex behaviors using workflows and durable objects, making it suitable for production-grade autonomous systems. Overall, Cloudflare Agents aims to streamline the development of scalable AI automation that operates close to users for improved performance. -
22
agents.md
A simple, open format for guiding coding agents
openai/agents.md is a repository whose primary file is AGENTS.md, a proposed open, lightweight convention (i.e. Markdown file) for guiding coding agents in software repositories. The idea is that AGENTS.md acts as a “README for agents”: a predictable, structured place where humans can put instructions, conventions, build/test commands, environment setup, and other guidance that generative agents (e.g. code-writing, code-assisting tools) should consult when operating in the repo. Instead of putting everything in README or doc files (which are more human-oriented and might mix high-level narrative), AGENTS.md is intended to surface agent-relevant details that help them “do the right thing” (tests, style, project structure, tooling). -
23
openclaw-kapso-whatsapp
Give your OpenClaw AI agent a WhatsApp number
openclaw-kapso-whatsapp is a plugin repository designed to extend the OpenClaw AI agent by giving it a dedicated WhatsApp phone number using the official Meta Cloud API via Kapso, enabling direct interaction through one of the most widely used messaging platforms. This integration allows the autonomous AI assistant to send and receive messages on WhatsApp, turning the agent into a real-world task performer accessible through text conversations. The plugin is built in Go and handles communication entirely through cloud APIs, avoiding the risk of bans that come with unofficial or reverse-engineered interfaces. Projects like this make it possible for OpenClaw users to automate tasks, interact with personal contacts, or provide AI-driven services without building a custom bot infrastructure from scratch. Because OpenClaw itself runs on the user’s own hardware and can access external services, this WhatsApp extension serves as a bridge between the AI agent and daily messaging workflows. -
24
/last30days
Claude Code skill that researches any topic across Reddit + X
/last30days is a specialized Claude Code skill designed to research current trends and practices across Reddit, X, and the wider web from the last 30 days, synthesize that data, and produce copy-paste-ready prompts or summaries that reflect what the community is actually talking about now. Rather than returning generic model responses, it intelligently analyzes social media and community discussions to identify what’s genuinely trending or working in practice across topics ranging from prompt techniques to tool usage or cultural trends. This makes it particularly useful for prompt engineers, content creators, and developers who want up-to-date prompts and insights that align with the most recent consensus and shared best practices in fast-moving fields like AI tooling. -
25
GitHub Agentic Workflows
GitHub Agentic Workflows
GitHub Agentic Workflows is an experimental CLI extension and framework for the gh GitHub CLI that lets developers author automation driven by natural language specifications instead of hand-written code, compiling those descriptions into GitHub Actions workflows that run AI agents (like Copilot, Claude Code, or Codex) on schedule or in response to repository events. By writing intent in markdown files, a developer can quickly generate .yml Actions workflows that perform tasks such as summarizing issues, automating triage, generating reports, or maintaining documentation, all without manually crafting YAML logic from scratch. The system emphasizes safety and guardrails, running agents in sandboxed environments with minimal permissions by default, and using “safe outputs” to constrain what the workflow can write back into the repository. It includes tooling for compiling, testing, and iterating on agentic workflows locally and integrates with GitHub’s existing Actions ecosystem.
