Install and Account

Download Embedr from embedr.app/download. You can also skip the install and use the web app at studio.embedr.app; see Embedr in the browser.

Installers

System	Files	Architectures
macOS	.dmg (and a .zip used for updates)	Apple Silicon (arm64) and Intel (x64), separate downloads
Windows	Embedr-Setup-<version>-x64.exe	x64
Linux	Embedr-<version>-x64.AppImage and Embedr-<version>-x64.deb	x64

macOS

1. Download the .dmg for your Mac (Apple Silicon or Intel).
2. Open it and drag Embedr onto the Applications link.
3. Start Embedr from Applications.
4. If macOS refuses to open it the first time, Control-click the app in Applications and choose Open, or allow it in System Settings → Privacy & Security.

On Apple Silicon, some Arduino toolchains are Intel-only. When a build needs Rosetta, Embedr says so and shows the command to install it:

softwareupdate --install-rosetta --agree-to-license

macOS asks for microphone access only when you open Voice Mode.

Windows

1. Run Embedr-Setup-<version>-x64.exe.
2. The installer is per user and does not need administrator rights. You can change the install folder; the default is %LOCALAPPDATA%\Programs\Embedr.
3. It creates Start menu and desktop shortcuts.
4. If Windows SmartScreen warns about the download, choose More info and then Run anyway if you trust the file you downloaded from embedr.app.

Uninstalling keeps your settings and projects.

If your board does not appear in the Device picker, install its USB serial driver. Common ones are CH340/CH341 (many Arduino-compatible boards), CP210x (many ESP32 boards) and FTDI.

Linux

• AppImage: make it executable (chmod +x Embedr-<version>-x64.AppImage) and run it.
• .deb: install it with your package manager, for example sudo apt install ./Embedr-<version>-x64.deb. The package installs into /opt/Embedr and registers Embedr as the handler for embedr:// links, which the desktop sign-in uses.

To flash boards and open serial ports, your user usually needs to be in the serial group (dialout on Debian and Ubuntu, uucp on Arch). Log out and back in after adding yourself:

sudo usermod -aG dialout "$USER"

On very large projects, Linux can run out of file watches (fs.inotify.max_user_watches). Embedr keeps running, but changes made outside the app in some folders may not show until you reopen the project. Raise the limit if that happens.

First launch

1. Sign in or create an account. The app opens on the sign-in screen.
2. Welcome to Embedr. Tick "I agree to the Terms of Service and Privacy Policy." The Session recording switch shares masked product interactions to help diagnose issues; you can turn it off here or later in Settings → Privacy. Click Continue.
3. Connect your AI accounts. Embedr includes free models, so this step is optional. Connect ChatGPT or OpenCode here to use your own subscription or balance, or click Skip for now. Click Start building when you are done.

You land on the home screen. See Start here for a first project.

Sign in

The sign-in screen offers Continue with Google or email and password (Or use email).

• Google on desktop opens your browser ("Opening browser for authentication..."). After you sign in there, the browser hands you back to the app through an embedr:// link. If your browser asks whether to open Embedr, allow it.
• Create account asks for Full Name, Email address and Password (at least 6 characters), and agreement to the terms. Disposable email addresses are refused.
• Verify your email. New email accounts get a verification link ("Check your inbox"). Until you click it, signing in shows "Please verify your email before logging in." and sends the link again. Use Resend verification email if it does not arrive.
• Forgot password? opens Reset password. Enter your email and click Send reset link.

To sign out on desktop, open Settings → Profile → Account Session → Sign Out.

Connect ChatGPT or OpenCode

You can connect your own model accounts during onboarding or later in Settings → AI Accounts.

Account	How to connect (onboarding; Settings uses Open ChatGPT and Save Key)	What you get
ChatGPT Plus / Pro	Click Connect. Embedr shows a code under "Enter code in your browser:". Click Copy code, then Open browser ↗, and enter the code on the OpenAI page.	A ChatGPT group in the model picker that uses your OpenAI subscription limits. Shows "Signed in as …" when connected.
OpenCode	Click Connect, paste your key into OpenCode API key (use Get API key to create one at opencode.ai), and click Connect OpenCode.	OpenCode Zen and OpenCode Go groups in the model picker, billed to your OpenCode balance.

Connections work on every plan, including Free. Until you connect, the model picker shows Connect ChatGPT and Connect OpenCode rows instead of those groups. Disconnect from the same place.

On desktop the connection is stored on your computer, in the app's data folder. Requests made with a connected account count against that account, not your Embedr plan. See Models.

Plans

Your plan decides which Embedr-provided models you can use and how much usage is included. Settings → Usage & billing shows the current plan.

Plan	What the app shows
Free	"Free tier includes 50 AI requests per 30-day cycle." Extra usage is not available.
Embedr Access	A 30-day pass with $25 of included credit. It includes GPT-6 Sol and GPT-6 Luna. When it ends, the account returns to Free.
Hobby	"$25 monthly AI compute credits with cloud compiler and standard models." Models marked Hobby+ in the picker need Hobby or Pro.
Pro	"$100 monthly AI compute credits with priority access and cloud compiler." Models marked Pro need Pro.

A locked model in the picker says "View plans to use this model".

Usage & billing

Section	Contents
Active Plan	"{Plan} Plan" and Upgrade Plan (Free) or Manage Subscription (paid).
Usage limits	"Resets {date}", Monthly plan limit with "{n}% remaining", Included in plan, Extra usage ("Not enabled on Free tier" or "Billed pay-as-you-go at the end of the billing cycle"), and Prepaid balance.
Billing & Invoices	Customer Portal opens billing and invoices in your browser.

The chat composer has a usage meter as well. It shows Usage remaining, the allowance (30-day allowance on Free, Monthly plan credit on paid plans) and "Plan usage resets {date}", with a button to upgrade, add credit or manage billing.

When you reach a limit

Message in the chat	What to do
Free Tier 30-Day Limit Reached: "You've reached your 50-request allowance for this 30-day period."	Click Upgrade to a Paid Plan, wait for the reset date, or switch to a connected ChatGPT or OpenCode model.
Embedr Access Credit Used	Click View plans. Your ChatGPT or OpenCode connection keeps working.
Hobby Credit Exhausted or Pro Credit Exhausted	Click Add Overage Credit, or wait for the monthly reset.

The Free allowance runs on a rolling 30-day cycle. The usage meter shows the date it resets.

Updates

Embedr checks for updates a few seconds after it starts and every few hours while it runs. Updates download in the background.

1. An update card appears at the bottom left: Update available, then Updating Embedr with a progress bar ("12.4 MB of 98.1 MB", percentage).
2. When the download finishes, the card says Update ready: "Restart Embedr to finish installing. Otherwise, it will update when you next quit."
3. Click Restart and update, or keep working and let it install the next time you quit. Minimize update shrinks the card to a small pill.

If the agent or a flash is still running when you click restart, Embedr asks first: "Embedr is still working on an active task. Restarting now will stop it. Restart and install the update anyway?"

If a download fails, the card shows Update failed with Retry.

When your version is too old to keep working with the service, the chat shows Update Required ("Update Embedr to version {version} or later to continue.") with an Update Embedr button. If the app cannot update itself, download the latest installer from embedr.app/download.

Windows installs the update silently when you restart. On Linux the AppImage and .deb builds use the same update feed.

Windows recovery

An interrupted update can leave a temporary copy of the app behind. If Embedr starts from that copy, it relaunches itself from the installed location. If the build tools are missing, a dialog titled Embedr needs to be repaired says "Embedr's build tools are missing, so projects cannot be created or built." Click Download Embedr and reinstall; your projects and settings are kept.

Settings

Open Settings from the gear in the title bar or with Cmd+, / Ctrl+,. The sections are:

Group	Sections
Personal	General, Profile, Appearance, Usage & billing, Notifications
Integrations	AI Accounts, Local tools (desktop only)
Coding	Editor, Command Approval (desktop only)
System & Privacy	Privacy, About

See Settings and privacy for each section.

What lives on your computer

What	Where
New projects	~/EmbedrProjects/<project-name>/ (%USERPROFILE%\EmbedrProjects\ on Windows)
Opened, imported and cloned folders	Wherever they already are. Embedr records them; it does not move them.
Project list and chat history	~/EmbedrProjects/.embedr/ (projects.json, and one folder per project with its threads)
ChatGPT and OpenCode connections, Git tokens you chose to remember, local settings	The app's data folder for your operating system

Your firmware, hardware sources and specs stay in the project folder. Sign-in, the agent, datasheet processing, model lists and updates need an internet connection, as do cloud tasks such as hosted KiCad checks when KiCad 10 is not installed locally (Settings → Local tools). Editing files, compiling with a toolchain that is already installed, flashing and the serial monitor work from local files.

Deleting a project from the home screen removes its chat history with it. Delete project also deletes the folder of a project Embedr created; Remove from Embedr leaves an opened folder on disk. See Project anatomy.