Subtext vs FrankenWhisper

Subtext is a communication tool built for autistic adults. Neurotypical messages are full of subtext: "no rush" that means hurry, "we should talk" that means nothing bad, "fine" that means anything. Subtext reads the message with you and shows what it likely means — never pretending there's only one answer. DECODE Paste a message (or share it straight from Messages, Mail, or any app). Subtext shows the plain meaning, then 2–3 possible readings with how likely each one is, the tone, and what the sender probably expects. Multiple readings, always — because certainty would be a lie. Screenshot a whole conversation and Subtext reads it in context. DRAFT Write what you actually want to say, bluntly. Subtext turns it into a reply in YOUR voice at the armor level you choose: direct, neutral, or polished. Your "no" never becomes a "maybe." Add samples of your writing so drafts sound like you, not like AI. TASKS Any overwhelming task becomes small concrete steps, sized to the energy you have. A quiet body-double mode keeps you company while you work. ENERGY Track your social battery in seconds, spot low-streaks, and send a pre-drafted "I need to reschedule" before burnout picks the words for you. REHEARSAL Practice a hard conversation before it happens — with an AI that plays the other side, shows how it might branch, and recaps only what you did well. BUILT DIFFERENT, ON PURPOSE • Nothing stored: decodes and drafts are processed, then gone. No account. No history on our side. • Never used to train AI. The switch doesn't exist. • No streaks, no guilt screens, no red alarms, no fake cheer. • Calm by default: no sounds, no vibrations, no surprise animations unless you turn them on. • Everything personal (people profiles, tasks, energy logs) stays on your phone, deletable in one tap. Subtext is a tool, not a therapist, not a medical device, and not trying to make you "more normal." It's for adults 18+. Autistic? Help build this. Our advisors are paid, credited, and hold veto power: thai@subtext.im

Turn recordings and videos into useful text without uploading them. FrankenWhisper transcribes microphone recordings, audio files, and playable videos from Photos, Files, or the share sheet. The full-quality lane uses the large-v3-turbo model. Live dictation and the optional custom keyboard use a smaller model for low latency, with local noise reduction in the real-time path. Speaker identification can separate up to four voices. Add names when you know them or leave them blank and use color alone. Export the transcript as text, Markdown, HTML, SRT, or JSON with timestamps and speaker structure. Subtitle Studio creates a new H.264 video with customizable captions, font, size, position, and color. Word-level timing drives a karaoke highlight, and named speakers can receive distinct labels and color schemes. All transcription, speaker analysis, denoising, timing, and video rendering run on the device. Model downloads are the only required network activity. The keyboard's Full Access setting is used only for the local App Group handoff between the keyboard and FrankenWhisper; the keyboard extension has no network client.