handup
One approval inbox for all your AI agents
handup is an approval inbox for AI coding agents. When an agent wants to do something consequential, like run a database migration, push a release or send an email, it sends a request with the real thing attached: the command, the diff, the file, the email draft or an image. You approve it, or deny it with feedback such as "use staging instead", from `handup inbox` in the terminal, the desktop app or the Android app. A denial reaches the agent as a normal answer, so it changes course instead of retrying.
`handup setup` connects the agents it finds: Claude Code and Codex get a permission hook plus MCP; Cursor, omp, Gemini CLI, opencode and other MCP clients use `handup mcp`; scripts and CI use `handup ask --wait`. Requests and decisions stay in a daemon on your own computer, with no handup account. handup is a review tool, not a sandbox.
Runs on Linux, macOS and Windows (beta) with an Android companion. 14 days free with every feature and no signup, then $30 one-time per person.
Key Features & Capabilities
Core highlights, automation workflows, and capabilities offered by this tool
Exact previews
commands, diffs, code, Markdown, JSON, HTML, images, video, audio, PDFs
Deny with feedback the agent reads as a normal answer
Works with Claude Code, Codex, Cursor, omp, Gemini CLI, opencode, any MCP client, shell/CI
Terminal inbox, desktop app, Android app (swipe to approve), web inbox
Local-first daemon; no account; optional self-hosted end-to-end encrypted relay
Desktop Run returns exit code so the agent doesn't run it twice
Software developers, DevOps engineers, AI-agent builders, security-conscious technical teams, and individual power users running autonomous coding or workflow agents.
How It Works (Step-by-Step)
The end-to-end user journey from input to automated output
Step 1
Connect handup to an MCP-compatible AI agent, Claude Code permission hook, or automation via its CLI/local HTTP API.
Step 2
When the agent needs approval, it sends an action request to handup containing the real diff, command, file, email draft, or other preview. The request is hash-bound to the exact content shown.
Step 3
Review the request in the desktop, web inbox, or Android app, then approve, deny with feedback, edit a draft, or answer a question. handup records the decision but never executes the action itself.
Step 4
The agent receives the decision and feedback; if denied, it can revise its plan and request approval again.
Frequently Asked Questions
Quick answers to common questions regarding functionality, billing, and access
Is handup free to try?
Who is handup for?
How does handup work?
Is my data private?
Which platforms are supported?
Who handles payments and refunds?
Need a Custom AI Solution Built for Your Business?
Turn your roadmap into production-grade systems — from custom LLMs and autonomous agents to complex workflow automation. Submit your project requirements and receive tailored technical blueprints and verified talent proposals.
Explore Live AI Calculators, Leaderboards & Side-by-Side Comparisons
LLM API Cost Calculator
Accurately estimate token consumption and operational costs across 50+ frontier LLMs including GPT-4o, Claude 3.5, Gemini, and DeepSeek.
AI Model Leaderboard
Evaluate frontier AI models side-by-side: reasoning benchmarks, code generation accuracy, latency scores, and price-to-performance rankings.
AI Comparison Matrix
Explore head-to-head architectural comparisons, pricing structures, feature breakdowns, and verified alternatives across leading AI software.
