AI Email Writer - Draft Emails in Your Browser

Fill in who the email is for, what it should achieve, and the key points, then press Draft email - the tool assembles a complete draft with a subject line, greeting, body, and sign-off. Unlike cloud email writers, everything runs inside your browser: the brief, and any email you paste in, never leave your device.

From bullet points to a finished draft

Most emails start as three or four loose thoughts: two colleagues are out sick, the deliverables are still on time, Monday morning works. Typing those as one-per-line key points is fast; turning them into a polite, well-ordered message is the slow part. The drafting form takes the recipient, the goal, and the points, and produces the full email in one step. Four tones - professional, friendly, short-and-direct, and apologetic - change the greeting, the opening line, and the sign-off, so the same brief can come out as a crisp status note or a careful apology.

The draft always ends with a [Your name] placeholder and a copy button. The tool never sends anything: you copy the finished text into Gmail, Outlook, or whatever mail app you already use, and send it from there.

Replying to an email you received

Switch to reply mode and paste the message you were sent. The draft opens by acknowledging their email, then works through your key points as the answer. Because the pasted email is processed on your device, this also works for messages you would not paste into a cloud service - an HR thread, a customer complaint, a supplier negotiation.

Three ways the draft gets written

The page works at three levels, and it tells you which one produced the draft on screen. With no download at all, a template engine builds the email deterministically from your brief: greeting from the recipient, one paragraph per key point, closing matched to the tone. If your browser ships built-in writing AI, the page uses it with nothing to download. And for more natural wording on any hardware with WebGPU, you can load a small open model (Llama 3.2 class) into the browser with one click - the download size is shown before it starts, and the files are cached so the next visit loads from disk. The same model cache is shared with the private AI chat tool, so a model downloaded on either page works on both.

Honest expectation: the on-device models are small, 1-3 billion parameters. They produce a solid first draft - read it before you send it, especially names, dates, and numbers. If you want an open-ended back-and-forth instead of a one-shot draft, that is the private AI chat page's job; if you already have a written text and only want its tone changed, the paraphrasing tool does exactly that.

← Back to Utility Tools