A personal AI agent for macOS. Turns intent into completed work, running from your menu bar.
Today's AI tools keep you hostage — Cursor locks you in a window, Claude Code blocks your terminal. Switch away, and you'll miss prompts or return to a stale [Y/n] that's been waiting for minutes.
Motive lives in your menu bar. AI works in the background. When it needs you, a lightweight popup drops down from the menu bar — no need to switch apps. One click, done, back to your flow.
| Desktop Apps | CLI Tools | Motive | |
|---|---|---|---|
| Where it lives | App window | Terminal | Menu bar |
| Permission prompts | Buried in UI | Blocks terminal | Menu bar popup |
| Switch apps? | Miss responses | Hangs silently | AI finds you |
- Intent-first — Describe tasks in natural language, press Enter, done. No conversation needed.
- True background execution — AI works as a background process, like compiling or rendering.
- Menu bar notifications — Permission requests drop down from the menu bar, not buried in an app window.
- Ambient status — A subtle menu bar icon shows progress without demanding attention.
- You are the final arbiter — Like macOS system permission dialogs, you approve only what matters.
- Local-first — All data stays on your machine. Only AI API requests leave your device.
- Native macOS — Built with SwiftUI and AppKit. No Electron, no web views.
- Multi-provider — Claude, OpenAI, Gemini, or fully local with Ollama.
motive_.mp4
Download the latest release for your architecture:
| Chip | Download |
|---|---|
| Apple Silicon | Motive-arm64.dmg |
| Intel | Motive-x86_64.dmg |
First launch: If blocked, go to System Settings → Privacy & Security → Click "Open Anyway"
- Click the menu bar icon → Settings
- Select your AI provider (Claude / OpenAI / Gemini / Ollama)
- Enter your API key
- Press
⌥Spaceto summon the command bar - Describe what you want done
- Press Enter — the bar disappears, you're free
- Check the menu bar icon for status; click to view details
| Shortcut | Action |
|---|---|
⌥Space |
Open command bar |
↵ |
Submit intent |
Esc |
Dismiss command bar |
⌘, |
Open settings |
git clone https://github.com/geezerrrr/motive.git
cd motive
open Motive.xcodeprojThe OpenCode binary is bundled automatically during release builds. For development, place it at Motive/Resources/opencode.
- macOS 15.0 (Sequoia) or later
- API key for Claude, OpenAI, Gemini, or local Ollama setup
How is this different from Cursor / Claude Desktop / Claude Code / Gemini CLI?
Desktop apps (Cursor, Claude Desktop) are window-locked — you must stay in their interface.
CLI tools (Claude Code, Gemini CLI) are terminal-locked — if you switch away, the terminal hangs waiting for input. You come back to find a cryptic [Y/n] prompt that's been sitting there for 20 minutes.
Motive lives in your menu bar — the AI runs as a background process. When it needs your input, a popup drops down from the menu bar. One click, done, back to your flow.
Think: desktop apps are like a colleague who insists you sit in their office. CLI tools are like someone who emails you but marks it "urgent" with no notification. Motive is like a colleague who taps your shoulder only when necessary, handles everything else autonomously.
Why does Motive need Accessibility permission?
To register the global hotkey (⌥Space) that summons the command bar from anywhere. Without it, you'd need to click the menu bar icon every time.
Is my data sent to the cloud?
Motive is local-first. Sessions and history stay on your machine. The only network traffic is API requests to your chosen AI provider. Use Ollama for 100% offline operation.
Can I use a local LLM?
Yes. Select Ollama as your provider and point it to your local instance. Zero cloud dependency.
What can Motive do?
Anything an AI coding agent can do: refactor code, generate files, run scripts, organize projects, write docs, and more. Motive passes your intent to OpenCode, which has full filesystem and terminal access.
- Multi-language UI — English, 简体中文, 日本語
- Browser automation — Full support for web scraping, form filling, and browser-based workflows
- Multi-task queue — Task queuing with parallel execution for independent tasks and sequential processing for dependent ones
- Task resume — Interrupt and resume long-running tasks, preserving state across app restarts
- Custom Skills — User-defined skills via
~/.motive/skills/directory, no code required - Personal Profile — Store personal context (name, preferences, work style) for more relevant AI responses
- Memory & RAG — Long-term memory with retrieval-augmented generation for context-aware assistance
- Task templates — Save and reuse common task patterns with customizable parameters
Powered by OpenCode — the open-source AI coding agent that makes autonomous task execution possible.
Let AI wait for you, so you don't have to wait for it.