No صفحه‌کلید swap, no separate برنامه to paste out of. Ramblr floats a single button over whatever you're already doing — Slack, Discord, a ویرایشگر کد, an email — و inserts شما words directly into the focused field the moment you're done talking. <ul> <li>Tap, ramble, tap. One button, everywhere. No mode switching, no separate transcript to copy-paste.</li> <li>کاملاً آفلاین if you want it. On-device transcription و on-device cleanup — nothing ever has to touch the network.</li> <li>Or route it through the best cloud models. OpenAI و Gemini برای transcription; OpenAI, Anthropic, و Gemini برای cleanup — chained in whatever fallback order you configure, شما own API keys, no relay server in between.</li> <li>Fastest of both worlds. Instant on-device transcription, cloud-grade cleanup — dictation feels instant و still reads like you actually meant to write it.</li> <li>Cleanup styles that match how you actually talk. Formal, Casual, یادداشت‌ها و lists, Email, Concise — or write شما own. Ramblr strips the "ums," collapses شما self-corrections, و keeps شما meaning, not just شما words.</li> <li>It learns شما vocabulary. Feed it شما project names و jargon once; cleanup stops mangling them forever.</li> </ul> Ramblr uses Android's Accessibility Service برای one narrow reason: to insert dictated text into the currently focused text field across برنامه‌ها. It does not replace شما صفحه‌کلید. It does not run background automation. It only acts after you explicitly tap the overlay button. <b>Five ways to start dictating:</b> the floating icon (default), the optional Ramblr Voice صفحه‌کلید, Android's سیستم accessibility button or a volume-key hold, a Quick تنظیمات tile, or by selecting existing text in any برنامه و tapping "Ramblr" to clean it up. Two of those — the voice صفحه‌کلید و the select-text integration — work با the Accessibility Service turned off entirely. Ramblr Voice is a voice-only input method, not a full replacement صفحه‌کلید: it has no letter, number, or symbol rows. <b>Network connections:</b> Ramblr makes no network connection on its own. Every connection it can make is listed here: <ul> <li><b>Speech و cleanup model downloads (on request).</b> On-device transcription needs a speech model, which is downloaded only when you pick one — from github.com (the sherpa-onnx project's release files) or huggingface.co. Optional on-device cleanup و optional silence auto-stop download their models the same way, only when you turn those features on. Nothing is downloaded until you ask برای it, و models are verified against a known SHA-256 before being installed.</li> <li><b>Optional cloud transcription و cleanup (off by default).</b> Only if you enable it و supply شما own API key. Requests go straight from گوشی شما to the provider you chose (OpenAI, Anthropic, Google Gemini, or a self-hosted OpenAI-compatible endpoint you point it at) — there is no Ramblr server in between.</li> <li><b>Cloud model list (only while setting up a cloud provider).</b> When you open the cloud provider setup screen, Ramblr refreshes its list of available cloud models — which models each provider offers و their prices — from a public file on gist.githubusercontent.com, at most once a day. This exists so a newly released cloud model can be offered بدون shipping an برنامه update. It is never fetched if you don't use cloud providers, it carries no identifying information, و a bundled list is used if the fetch fails.</li> </ul> There is no analytics, no telemetry, no crash reporting, و بدون تبلیغاتvertising of any kind. This F-Droid build also contains no update checker: it never مخاطبین GitHub to look برای جدید versions, و updates come from F-Droid like any other برنامه. <b>Privacy:</b> combine local transcription با local cleanup و nothing ever leaves the device. There is no backend برای this برنامه — in every cloud mode, requests go straight from شما phone to the provider using شما own API key, با no relay server in the middle. Cl