AI Agents
ai-berkshire
AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code /…
★ 16.6k starsMIT
View guide →Local-first agentic workspace where Claude Code, Codex and a built-in agent share memory, apps and tools.
holaOS is an open-source ai agents project by holaboss-ai (holaboss-ai/holaOS on GitHub). It has 11,454 GitHub stars and 728 forks, is released under the Modified Apache-2.0 license, and was last updated on GitHub on 21 August 2026.
Choose holaOS when your files, memory and connected accounts should stay on your own machines while agents work across them.
Use the one-line installer from the README, or clone the repo and set up the desktop app from source.
Sign in to integrations like Gmail, Notion, Slack, GitHub or Linear with one-click OAuth, plug in any MCP server, and either use the built-in models or add your own provider keys.
Ask in plain language - a report, a spreadsheet, follow-ups across your apps - and approve each action as the agent assembles a real deliverable file.
Claude Code, Codex and the built-in holaOS agent all run in the same workspace and share the same memory, tools, skills and apps.
One memory, stored on your disk as plain, readable and editable files, is shared by every connected agent instead of being siloed per tool.
Apps from the in-workspace marketplace open as real interactive surfaces next to the agent chat; you can also bring your own app via a URL plus an MCP server.
One-click OAuth for 50+ integrations such as Gmail, Notion, Slack, GitHub and Linear, plus connections from Slack, Feishu, DingTalk and WeChat with per-tool approval.
Plug in any MCP server, install community servers in one click, and bundle skills into reusable Combos; HolaHub recipes package model, prompt, skills and integrations for one-click install.
Agents produce finished .xlsx, .pptx and .docx files, can browse the web signed-in as you, generate image/video/audio, and run scheduled or triggered automations - with every action approved by you.
Get the holaOS source locally; most users can instead use the one-line installer from the README.
git clone https://github.com/holaboss-ai/holaOS.git
cd holaOSInstall dependencies and copy the example environment file, then prepare the local runtime.
npm run desktop:install
cp apps/desktop/.env.example apps/desktop/.env
npm run desktop:prepare-runtime:localStart the desktop app and sign in or add your own model keys.
npm run desktop:devThe holaboss-ai team maintains it as an open-source project, with a hosted account and enterprise tier offered alongside the self-run desktop app.
The desktop app is open source under a modified Apache 2.0 license with extra conditions on commercial distribution and branding; built-in model access and enterprise features are part of its paid offerings.
No. It ships with built-in frontier models usable through a single holaOS account with no keys to manage; bring-your-own-key for OpenAI, Anthropic or any compatible endpoint is optional.
Those are single coding agents in a terminal; holaOS is a whole workspace that runs several agents over your apps, files and browser with one shared local memory and approval controls.
Run the one-line installer from the README (or clone and use the npm desktop commands), connect one integration, and give the agent a small task.
AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code /…
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
A local-first AI agent workspace for coding, writing, design, research and automation.
Community-built terminal coding agent: bring your own model, keep code local.
Bring Claude Code, Codex, and Gemini CLI into Obsidian via Agent Client Protocol.
Orchestrate Claude Code, Kiro, and Codex together in isolated tmux sessions.
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if holaOS is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.