Offline Speech to Text / Offline Speech to Text is on Nostr

Private, fully offline speech to text.

Last Notes

2026-10-30 08:18:05 UTC

Offline Speech to Text 0.23.0 is available on Zapstore. Keyboard dictation can keep going when the keyboard hides or the screen turns off, as long as the app's notifications and dictation channel are enabled. If you switch fields while recording, your text goes to private recovery. On Android 13+, optional Bluetooth mic support (off by default) now prefers the headset selected for media playback. [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-29 10:13:35 UTC

Offline Speech to Text 0.22.0 is released: file transcription without an interactive recording-duration limit (dictation stays bounded), audio/video sharing with Copy/Share/Save, an optional Bluetooth headset or LE hearing-aid mic on Android 12+, and reliability fixes. Updates keep your models and settings. [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-25 09:33:45 UTC

Offline Speech to Text 0.21.0 is out. It adds the Canary 1B v2 Q4 model and Languages I speak, which lets supported models detect among up to eight chosen languages using a separately downloaded, verified detector. English filler cleanup is now multilingual-safe, and Donate in Settings → About opens a Lightning wallet to the app's own address. Recognition stays on-device and offline after setup. [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-21 11:38:02 UTC

Offline Speech to Text 0.20.0: AMOLED amber dark theme, a setup-focused Home screen, a clearer Speech models view, 10 tested download choices and an optional Parakeet Q8 download. Works offline once a model is set up. [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-21 09:00:46 UTC

Offline Speech To Text 0.19.0 is out. Setup now starts with a speech model and microphone permission. The model screen shows one recommendation for your device and language, and the dictionary separates Preferred words from heard→written Corrections. There's also a true-black dark mode for AMOLED screens. Recognition stays on your device. [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-20 12:18:11 UTC

Offline Speech to Text 0.18.0 is on Zapstore. English dictation now removes standalone "uh" and "um". With multilingual Parakeet, select English to get this cleanup. Preferred names with letters and digits keep your spelling, TalkBack announces keyboard status changes, and there's a new microphone icon. Recognition and cleanup run offline on your device. https://zapstore.dev/apps/app.offlinespeechtotext

2026-10-18 13:27:08 UTC

Offline Speech to Text 0.17.0 is available.
Detailed troubleshooting now offers 1 hour, 24 hours, Forever, then Cancel. Logging stays off by default and requires explicit consent. Forever keeps collection active until you turn it off; the local log still has a size limit. Audio and transcripts remain excluded.
This update also improves the warning’s readability and simplifies its controls. The speech engine is unchanged.
Download: https://zapstore.dev/apps/app.offlinespeechtotext

2026-10-18 12:56:16 UTC

- reply

Okay, that makes sense. By the way, we just put out another release with big improvements!

2026-10-18 12:41:24 UTC

- reply

Cool idea! Are you using the speech engine OSTT exposes to other Android apps, so Roadstr can reuse its transcription model and avoid a separate download?

2026-10-18 12:40:44 UTC

Offline Speech to Text 0.16.0 is out.
This release brings a newer offline speech engine, streaming audio-file transcription with progress and partial-text recovery, and safer keyboard insertion. The APK is about 40% smaller than the previous public release. Custom dictionaries and configurable pause detection remain available; early audio preparation stays off by default.
Download: https://zapstore.dev/apps/app.offlinespeechtotext
Source: https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/offline-speech-to-text

2026-10-06 12:56:30 UTC

Offline Speech to Text 0.14.0 adds explicit one-hour or 24-hour troubleshooting,
corrects event timestamps and callback-delivery logs, and reuses the microphone
meter animation. Diagnostics stay opt-in and exclude audio and message text.
[Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext) · [Source](https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/relay.ngit.dev/offline-speech-to-text)

2026-10-06 08:39:41 UTC

Offline Speech to Text 0.13.1 simplifies diagnostic validation and marker writes. Recognition, models and privacy defaults keep their existing behavior. It includes optional early processing for long keyboard dictation and one-hour troubleshooting from 0.13.0. Audio and transcripts stay outside diagnostics.
[Update from Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext)

2026-10-05 17:53:23 UTC

Offline Speech to Text 0.13.0 adds optional early processing for long keyboard dictation and separately warned, one-hour troubleshooting. Preparation uses existing overlapping windows and retains the recording; one final result follows Stop. Audio and transcripts stay out of diagnostics.
https://zapstore.dev/apps/app.offlinespeechtotext

2026-10-04 09:12:41 UTC

Offline Speech to Text 0.12.2 drains buffered audio before transcription when apps request Stop, reports input failures explicitly, and improves private, opt-in diagnostics.
Download: https://zapstore.dev/apps/app.offlinespeechtotext
Source: https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/relay.ngit.dev/offline-speech-to-text
Release notes: https://njump.me/naddr1qqwk7enxd35kuefdwdcx2etrdqkhgmedw3jhsapdxqhrzv3wxgqs6amnwvaz7tmwdaejumr0dsq3yamnwvaz7tm0venxx6rpd9hzuur4vgpzpys5pkhzxd9dqp4ger8du6p5f6y43tcnzqktjzmwvahq5vumtay4qvzqqqr4guhy7pwd

2026-10-01 08:49:52 UTC

OSTT 0.12.1 is out: simpler dictionary cleanup and clearer opt-in diagnostics, including build markers and content-free recognition error codes. Existing matching and speech behavior stay the same.
Update: https://zapstore.dev/apps/app.offlinespeechtotext
Source: https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/relay.ngit.dev/offline-speech-to-text
Verified APK: https://blossom.primal.net/7f7f5a13b9ec5c4f28491b79bec2857be8982e1655d55968615d95f7bcace005

2026-09-30 22:06:37 UTC

Offline Speech to Text 0.12.0 is out. The personal dictionary handles exact compound names and protects quoted and linked text. Auto-stop users can choose a two, three or five second pause. The update keeps installed models and settings.
[Release notes](https://njump.me/naddr1qvzqqqr4gupzpys5pkhzxd9dqp4ger8du6p5f6y43tcnzqktjzmwvahq5vumtay4qyxhwumn8ghj7mn0wvhxcmmvqyf8wumn8ghj7mmxve3ksctfdch8qatzqqwk7enxd35kuefdwdcx2etrdqkhgmedw3jhsapdxqhrzv3wxq2pdexr) · [Zapstore](https://zapstore.dev/apps/app.offlinespeechtotext) · [Source](https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/relay.ngit.dev/offline-speech-to-text) · [APK](https://blossom.primal.net/9ea63d0450bbed54e0050486c3dd7cc2e9b26d669c1b613aa8ccd22645cca485)

2026-09-11 17:02:46 UTC

- reply

Offline Speech to Text 0.1.19 is out: private, fully offline speech-to-text for Android. This initial release fixes caller-supplied audio in the background recognition service.
#naddr1qq…enay
Source: https://gitworkshop.dev/npub1jg2qmt3rxjksq65v3nk7dq6yaz2c4uf3qt9epdhxwms2xwd47j2s38a228/offline-speech-to-text