The open-source desktop automation platform that combines a visual builder, real browser control, AI agents, and 80+ API integrations — all running locally on your machine.

Download the latest release for your OS from Releases.

Or build from source:

git clone https://github.com/Dyan-Dev/loopi.git
cd loopi
pnpm install
pnpm start

Build agentic workflows that connect AI to real browser actions and 80+ APIs — without sending your data to the cloud.

Loopi gives AI models the ability to act: browse the web, call APIs, query databases, send messages, and process data — all orchestrated visually. Think of it as giving your AI hands and eyes.

Connect any LLM:

• OpenAI (GPT-4o, GPT-4o-mini) — cloud
• Anthropic (Claude) — cloud
• Ollama (Llama, Mistral, Gemma, Phi, etc.) — runs 100% locally, your data never leaves your machine

What you can build:

• AI agents that browse websites, extract data, and make decisions
• Workflows that use LLMs to summarize, classify, or generate content — then act on the results
• Chains that call an API, feed the response to an AI model, and post the output to Slack/Discord/Notion
• Local-first AI pipelines with Ollama — no API keys, no cloud, no cost
• AI copilot that generates, explains, and auto-fixes your workflows

Why this matters: Most AI agent frameworks are code-only. Most no-code tools are cloud-only. Loopi is both visual and local-first, with the option to go cloud when you want to. You own your data, your models, and your workflows.

Loopi is the only tool that gives you visual workflows + real browser automation + 80+ API integrations + local & cloud AI + typed variables, all local-first and open source.

Visual Workflow Builder — Drag-and-drop node editor powered by ReactFlow. Build complex automations without writing code.

Real Browser Automation — Navigate, click, type, extract data, upload files, take screenshots — all in a real Chromium window you can watch.

80+ Integrations — Connect to services out of the box:

Typed Variable System — Auto-detected types with dot notation and array indexing: {{user.name}}, {{items[0].price}}, {{apiResponse.data}}.

Data Transforms — JSON parse/stringify, math operations, string operations, date/time, filter arrays, map arrays, and inline code execution.

Conditional Logic & Loops — Branch workflows with conditions, iterate over arrays with forEach loops.

Credentials Manager — Store API keys and tokens securely. Select credentials from a dropdown when configuring steps.

Scheduling — Run automations on intervals, cron expressions, or one-time schedules.

Import/Export — Save and share automations as JSON. Includes example workflows to get started.

Electron, React 19, TypeScript, ReactFlow, Tailwind CSS, Radix UI, Biome

• Getting Started — Installation and first automation
• Steps Reference — All step types and their fields
• Variables — Variable system and access patterns
• Credentials — Managing API credentials
• Architecture — System design and data flow
• Component Guide — React component structure
• Adding New Steps — How to add step types
• Development Workflows — Common dev tasks

We welcome contributions — especially around AI agent capabilities, new integrations, and workflow templates.

pnpm install        # Install dependencies
pnpm start          # Run in development
pnpm run format     # Format with Biome
pnpm test           # Run tests
pnpm run make       # Build for current platform

Ideas for contributors:

• Add new AI model providers (Gemini, Cohere, local GGUF models)
• Build agentic workflow templates (research agents, monitoring agents, content pipelines)
• Add new service integrations
• Improve the AI copilot (auto-generate workflows from natural language)
• Add tool-use / function-calling support for connected LLMs

See CONTRIBUTING.md for guidelines. This project uses Biome for formatting and linting.

• GitHub Issues — Bug reports and feature requests
• GitHub Discussions — Questions and community
• Email: [email protected]

O'Saasy License