Need to send a letter, fire off a cover letter, or finish a short document and only have your phone? Dictate it. AgentDoc turns what you say into a properly formatted document – headings, paragraphs, even a clean PDF – without you ever touching the on-screen keyboard. No app to install. No autocorrect ruining your words.
Smartphones replaced laptops for almost everything – except writing a letter or short document. Four hard constraints make mobile document writing unpleasant to the point of avoidance.
An A4 page at readable font size doesn't fit a 6-inch screen. You either zoom in and lose the big picture, or zoom out and can't read the text. There is no middle.
Pull up the on-screen keyboard and half the document disappears behind it. You're typing a formal letter on a keyboard that autocorrects "Sehr geehrte Frau Müller" into something embarrassing – and you can't see what you've written.
Try dropping the cursor between two specific characters on mobile. The magnifying loupe fights you. Selection handles drift. Three taps later you're editing the wrong line.
Desktop editors hide formatting behind nested toolbars and right-click menus that don't translate to touch. On mobile you get a stripped-down subset – or nothing.
If you don't type on mobile, every one of the four problems above disappears. The document stays full-screen, cursor placement becomes the agent's job, and every formatting operation – no matter how deeply buried on a desktop – is just a spoken sentence away.
Your phone shows the complete page the entire time. A single microphone button anchors the bottom corner. Everything else – headings, bold, colors, fonts, page breaks, pagination, PDF export – happens through voice commands that an AI agent translates into precise, typed operations on the document.
Because the agent operates on the document's structure (not your thumb's approximate position), precision is absolute. "Make the second revenue number green" lands on the right word every time, whether it's on page one or page seventeen.
Mobile interaction surface stays minimal on purpose: tap to open, hold to speak, swipe to page. Everything else is voice.
One button, always in the bottom-right. Tap it to start a voice session. The keyboard never appears – the document keeps the full screen.
"Add a new heading called Conclusions." "Make paragraph three italic." "Insert a page break before the summary." Phrase it however you would to a colleague.
The agent translates speech into typed tool calls that mutate the document directly. The page re-renders in real time – and the agent confirms each edit out loud.
Voice-first mobile editing isn't a compromise – for everyday short documents it's faster than a laptop, because you skip the entire setup of opening one.
Dictate a formal letter – opening, body, closing, signature – and export it as PDF without ever touching the keyboard. Format like "make 'Sehr geehrte Frau Müller' bold" works mid-sentence.
Dictate a cover letter on the train, refine paragraphs by voice, export PDF before you arrive. No laptop, no tray table, no compromises on formatting.
Got sent a proposal or contract and need to add a paragraph? Open it on your phone, dictate the addition, and the agent inserts it at the right place – no laptop required.
Standing in line, on the train, between meetings. One button, one mic, a finished short document by the time you arrive. No app to install – works in any mobile browser.
RSI, motor impairments, temporary injuries, or simply tired thumbs – dictate the whole letter and let the agent handle layout, paragraphs, and PDF export.
The kind of letter you keep meaning to write, but never sit down at a laptop for. Dictate it from the couch, send the PDF, done in five minutes.
Yes – that's exactly what AgentDoc is built for. Dictate the salutation, body paragraphs, and closing; tell the agent to bold the recipient's name or change the font; export a clean PDF. No keyboard, no autocorrect, no fighting the on-screen layout.
No. AgentDoc runs in any modern mobile browser – Safari on iOS, Chrome on Android. There is no app-store install, no subscription, and no client-side setup beyond creating an account and allowing microphone access.
A better keyboard still consumes half the screen, still makes you tap one letter at a time, and still doesn't solve cursor placement or formatting menus. Voice removes all four mobile problems at once – you get the full page, precise targeting via an agent, and access to every formatting operation with no menu at all.
AgentDoc uses Google Gemini Live and runs each instruction through typed tool calls against the actual document state. The agent must find the exact text, apply the change, and report what it did before the next instruction begins. This is significantly more reliable than open-loop dictation because every step is grounded in the real document – not an approximation of what was said.
Voice data is streamed to Google Gemini for processing. For sensitive or confidential documents, use the text chat mode instead – the same agent, the same tool suite, typed input, no audio stream. Both modes run over HTTPS with JWT-authenticated, per-user document isolation.
Yes. Say "export this document as a PDF" and the browser downloads a pixel-perfect A4 PDF. The same format you would get from the desktop editor – exportable and shareable directly from your phone.
Those tools transcribe speech into text at the current cursor position. They don't format, they don't restructure, they don't navigate, and they can't operate on parts of the document the cursor isn't in. AgentDoc is a document editor driven by voice – it understands the structure of the document and can edit any part of it by name.
Three reasons stack up. The screen is too small to show a full A4 page legibly. The on-screen keyboard covers half of whatever fits. And touch cursor placement is imprecise on small text – three taps just to land between two specific characters. Voice removes all three at once: the page stays full-screen, no keyboard appears, and cursor placement becomes the agent's job, not your thumb's.
Yes. After you're done dictating, say "export this as PDF" and the browser downloads a pixel-perfect A4 PDF you can attach to an email or share directly. Same format as a desktop export, no laptop needed.