OpenOri is a desktop AI agent that runs entirely on your machine. It remembers your preferences, understands your projects, controls your screen, runs tasks on autopilot, writes and previews files in real time, and picks the right tools for every request. It’s not a chat wrapper. It’s a personal AI that gets smarter every time you use it.Documentation Index
Fetch the complete documentation index at: https://docs.openori.dev/llms.txt
Use this file to discover all available pages before exploring further.
OpenOri is open source and free. All data stays on your machine. View on GitHub →
Get started in 2 minutes
Download OpenOri
Grab the latest release for your platform from GitHub Releases.
Add your API key
The onboarding wizard walks you through connecting Anthropic, OpenAI, Google, or Ollama.
Why OpenOri
Memory that persists
OpenOri extracts facts, preferences, and patterns from every conversation — and remembers them forever. Day 1 you tell it you prefer TypeScript strict mode. Day 100 it still knows.
Controls your computer
OpenOri sees your screen, clicks buttons, types text, scrolls pages, and presses keys. Automate anything visual — from filling forms to navigating apps.
Your keys, your data
100% local. Zero telemetry. API keys stored in your OS Keychain. Conversations, memories, and settings never leave your machine.
Community skills
Browse and install skill packs from the community — email drafting, data analysis, git workflows. One click to install, instant routing.
Reminders & scheduled tasks
Quick reminders that fire on time, recurring alerts, and autonomous agent jobs — morning briefings, PR digests, competitive monitoring, and more.
Embedded terminal
A full terminal lives inside OpenOri — multiple tabs, live output during tool execution, and drag-to-resize. No context switching.
Live file preview
OpenOri proposes file changes as reviewable artifacts. Preview HTML files live in a sandboxed iframe, render SVGs inline, and view rendered Markdown — all in-app.
Ambient intelligence
OpenOri works silently in the background — monitoring your services, learning your preferences, and surfacing what matters without you asking.
What you can do with OpenOri
| Say this… | OpenOri does this… |
|---|---|
| ”Remind me to call Sarah in 30 minutes” | Creates a reminder that fires a toast + OS notification |
| ”What reminders do I have?” | Lists all pending reminders with times |
| ”Every morning at 9am, summarize my GitHub PRs” | Creates a scheduled task that runs autonomously |
| ”Remember that I prefer dark mode everywhere” | Saves to persistent memory — recalled in future conversations |
| ”Take a screenshot and tell me what you see” | Captures your screen and describes the content |
| ”Plan out migrating our API to Hono” | Creates a multi-phase plan with tracked progress |
| ”Refactor this function to use async/await” | Reads your code, writes changes, shows a diff preview |
| ”Create an HTML page with a demo” | Writes the file and shows a live sandboxed preview |
| ”Search the web for the best TypeScript frameworks” | Runs a web search and summarizes the results |
| ”Install the email drafting skill” | Downloads and activates the skill from the community registry |
Explore
Installation
macOS, Windows, Linux, or from source.
Your first conversation
A hands-on walkthrough of what OpenOri can do.
Coding with OpenOri
Read files, run commands, write code, manage Git.
Research with OpenOri
Search the web, crawl pages, synthesize reports.
Automating workflows
Scheduled tasks, connectors, and recurring jobs.
All features
The complete list of everything OpenOri ships with.