What is Autohand Code?
Autohand Code is an AI coding workspace that can understand repositories, plan and make changes, run tools, and report evidence across Web, terminal, editor, mobile, and SDK surfaces.
Getting Started
Start with Autohand
Autohand Code is an AI coding workspace for the browser, terminal, editor, and phone. The quickest path is Autohand Dev in your browser: no local installation or provider key is required to begin. Choose a repository, describe the outcome, and review the work before accepting changes.
New to Autohand? Start on the Web. You can move to the CLI, an editor, or mobile later without changing the job you are trying to do.
Fastest path
Run a useful task in the browser
-
Open Autohand Dev
Go to dev.autohand.ai/new and sign in. This browser surface is the easiest first experience because there is nothing to install locally.
-
Choose a repository
Select the repository you want Autohand to understand. For a first run, use a project where you can review changes and revert them safely.
-
Ask for understanding before edits
Give the agent one bounded outcome. A read-only exploration is a good first task because it shows how Autohand reasons about your code without asking it to modify anything.
Inspect this repository without changing files. Explain what it does, identify the main entry points, and suggest one small improvement I can review.
Completion checkpoint
You should have a repository summary, named entry points, and one specific next step. If you do, your first Autohand task is complete.
Choose by outcome
What do you want to do?
- Understand or change a repository Start in Autohand Dev with no install, then move to another surface when you need it. Open Dev
- Work inside a local project Use the CLI for terminal-first coding, local tools, scripts, and repository workflows. Set up the CLI
- Stay in your editor Use VS Code, Zed, or JetBrains when a visual code-and-diff loop fits your work. Choose an editor
- Build an agent workflow Use the Code Agent SDK to embed prompts, events, approvals, and results in your own system. Open the SDK
Choose by work style
Start on the surface that fits
Every surface is an entry point to Autohand Code. You do not need to learn them all before you begin.
| Surface | Best for | Setup | Start here |
|---|---|---|---|
| Web Autohand Dev | The fastest first task, remote repositories, and browser-based work. | Sign in; no local install. | Open Dev |
| Terminal Autohand CLI | Local codebases, shell tools, automation, and terminal-first workflows. | Install the CLI, then connect a hosted or local model. | CLI guide |
| Editor VS Code, Zed, JetBrains | Working beside your code with an editor-native review loop. | Install the extension or plugin for your editor. | Editor guides |
| Mobile Autohand Dev or handoff | Starting, checking, or continuing work away from your computer. | Open Dev in your mobile browser or pair the iOS app with a CLI session. | Mobile options |
| SDK Code Agent SDK | Products, internal tools, CI, and programmable agent workflows. | Create an API key and install the SDK for your language. | SDK quickstart |
Use Autohand Dev like a mobile app
Add it to your Home Screen
On iPhone or iPad, open Autohand Dev in Safari, tap Share, then Add to Home Screen. On Android, open it in Chrome, open the browser menu, then choose Add to Home screen or Install app when that option is available.
Stay in control
Turn the first answer into safe progress
After Autohand explains the repository, ask it to plan one small change. Keep the request scoped, name the verification you expect, and inspect the diff before accepting it.
Propose a plan for the smallest useful improvement. Do not edit yet. Name the files you would change, the tests you would run, and any risk I should review.- Check the scope. The plan should match the outcome you asked for.
- Check the evidence. Ask for tests, build results, or runtime proof appropriate to the change.
- Check the diff. Review changed files before you commit, merge, or deploy anything.
Plans
Which Autohand plan should you use?
Start with Free if you are evaluating Autohand. Move to Pro or Max when you need more hosted-model capacity, or use your own provider key without consuming hosted credits.
Free
$0. Best for learning and occasional use. Includes 20 hosted credits per 5 hours, up to 200 per month.
Pro
US$20/month. Best for regular individual use. Includes 250 credits per 5 hours.
Max
US$200/month. Best for intensive daily work. Includes 1,000 credits per 5 hours.
Team
US$59/seat/month. Best for shared administration and team usage, with per-member limits.
Enterprise
Contact sales. Best for organization-wide controls, support, and custom requirements.
Bring-your-own-key usage is available on every plan and does not consume Autohand hosted credits. Hosted models through the SDK or API require a paid plan.
Quick answers
Autohand getting started FAQ
Can I try Autohand without installing anything?
Yes. Open Autohand Dev at dev.autohand.ai/new, sign in, select a repository, and start a task in your browser. The CLI and editor extensions are optional surfaces for local workflows.
Which surface should a new user choose?
Start with Autohand Dev on the Web for the shortest setup path. Choose the CLI for local terminal work, an editor extension for an editor-native review loop, or the SDK when you are building Autohand into another system.
Can I use Autohand on my phone?
Yes. Use Autohand Dev in a mobile browser and add it to your Home Screen for quick access. The iOS app can also pair with a running CLI session for mobile handoff.
Which plan is best for a first project?
The Free plan is the best default for evaluation and occasional use. Pro fits regular individual use, Max fits intensive use, and Team or Enterprise adds organization-oriented capacity and controls.
Do I need an AI provider API key?
Not to begin in Autohand Dev with hosted models. For CLI workflows, you can use Autohand hosted models where your plan permits, connect a supported provider key, or configure a local model.