In This Article
Speech-to-text on Android in 2026 is finally good enough to use as a primary text input for most users. The on-device speech recognition built into Gboard (Google’s keyboard) handles dictation in 70+ languages with near-court-reporter accuracy on a quiet input. The third-party tier (Whisper-based apps, Otter, Rev Voice Recorder) covers the deeper use cases.
This guide covers six speech-to-text apps that earn their slot in 2026: Gboard Voice Typing as the universal default, Whisper-powered Lupin, Whispr or Otter (depending on offline vs cloud preference), Otter for meeting transcripts, Rev Voice Recorder for journalism-grade transcription, and Live Transcribe for accessibility use cases.
What changed in 2026: Gboard added a 99-language offline pack (up from 13 in 2023), Whisper Large v3 became the on-device speech model for most premium apps, Otter rolled out a 2025 redesign with real-time speaker identification, and Live Transcribe (Google’s accessibility tool) added punctuation-aware transcription that brought it to parity with the paid tier.
TL;DR
Daily default: Gboard Voice Typing. Free, on-device for 99 languages, accurate enough for most prose.
Premium accuracy: Whisper-powered app like Lupin or Whispr. Cloud or on-device depending on phone hardware.
Meetings: Otter for real-time transcripts with speaker ID.
Gboard Voice Typing for the universal default
Gboard Voice Typing is built into Android and good enough for 80 percent of speech-to-text use cases. Tap the microphone icon on the Gboard keyboard (or hold the spacebar on some Android configurations), speak, watch the text appear. Free, no account, 70+ languages, offline support for the majority of those languages.
Punctuation: speak the punctuation aloud (‘comma’, ‘period’, ‘new line’). Auto-punctuation is on by default and handles most commas and periods automatically; for question marks and exclamation points, speak them explicitly.
Offline support: the on-device language packs ship with Gboard but you can also download additional packs in Settings, System, Languages and input, Voice typing, Offline speech recognition. The 99-language offline pack covers most users; cloud-only is the fallback for rare languages.
Whisper-powered apps for the accuracy bump
Whisper (OpenAI, open-source) is the most-accurate consumer speech-recognition model available. Multiple Android apps wrap Whisper into a friendly interface; the leaders in 2026 are Lupin (5.99 USD per month) and Whispr (free with optional 3.99 USD tip).
Whisper’s advantage over Gboard: better accent handling, better technical-term recognition (medical, legal, scientific vocabulary), better punctuation, and better handling of fast or accented speech. The trade-off is that most Whisper apps run in the cloud (your audio is processed off-device), which has privacy implications for sensitive transcription.
For on-device Whisper, Whispr supports the smaller (lower-accuracy) Whisper variants offline; the full Large v3 model requires a phone with at least 8 GB of RAM and runs only on more recent flagships. If your phone qualifies, on-device Whisper is the best of both worlds: top accuracy and full privacy.
Otter and Rev for meetings and journalism
Otter (free or 16.99 USD per month Pro) is the gold standard for live meeting transcripts. It generates speaker-identified transcripts in real time during Zoom, Google Meet, and Microsoft Teams calls; AI-summarized notes at the end of each meeting; highlight extraction; integration with Slack and Notion. Free tier covers 300 minutes per month; Pro lifts the cap.
Rev Voice Recorder (free, with per-minute paid transcription) is the Rev company’s mobile app. Record audio on the phone, optionally submit for human-verified transcription at 1.50 USD per minute. The human transcripts are court-reportable accuracy; the AI transcripts (free in-app) are at Whisper-Large quality.
For everyday users, Otter is the right pick because most use cases are meeting transcripts. For occasional journalism or legal contexts where verbatim accuracy matters, Rev is the right pick because the human-verified path is the only one that meets professional standards.
Live Transcribe and accessibility use cases
Live Transcribe (Google, free, preinstalled or downloadable on most Androids) is Google’s accessibility app for real-time speech-to-text. Originally designed for deaf and hard-of-hearing users; useful for anyone who wants live captions of an in-person conversation. The 2024 update added punctuation and the 2025 update added automatic language detection.
Live Transcribe also has a sound-notification mode that alerts you to specific sounds in the environment (doorbell, baby crying, fire alarm). For accessibility users this is genuinely life-changing; for general users it is a useful background tool to have installed.
Sound Amplifier (also Google, free) pairs with Live Transcribe for hearing-augmented use cases. Sound Amplifier boosts and clarifies live audio through headphones; Live Transcribe transcribes simultaneously. The combination is essentially a real-time captioned hearing aid.
For the broader question of underused Android accessibility and productivity tools, the BFA piece on most useful Android apps covers adjacent picks.
Quick take
For 80 percent of cases, Gboard Voice Typing is enough. Free, built-in, offline-capable.
For meeting transcripts, Otter. For maximum accuracy, a Whisper-powered app. For accessibility, Live Transcribe.
At a glance
| App | Cost | Best for | Offline | Accuracy |
|---|---|---|---|---|
| Gboard Voice Typing | Free | Daily dictation | Yes (99 langs) | Very good |
| Lupin (Whisper) | $5.99/mo | Premium dictation, transcripts | Partial | Excellent |
| Whispr (Whisper) | Free + $3.99 tip | Free Whisper alternative | On capable phones | Excellent |
| Otter | Free or $16.99/mo Pro | Meeting transcripts | No | Excellent (speaker ID) |
| Rev Voice Recorder | Free + $1.50/min human | Court-reportable accuracy | No | Pro-grade (human) |
| Live Transcribe | Free | Accessibility, real-time captions | Yes | Excellent |
FAQ
Is Gboard’s Voice Typing accurate enough for professional dictation?
For most non-specialized prose, yes. Email, notes, chat messages, basic memos all transcribe accurately. For specialized vocabulary (medical, legal, technical jargon), Whisper-powered apps still have an edge. For court-reportable accuracy, Rev human transcription remains the only reliable path.
Are these apps secure for sensitive content?
Gboard Voice Typing offline mode processes everything on-device with no cloud roundtrip; this is the most secure option. Whisper apps that run on-device are similarly secure. Cloud-based services (Otter, Rev, cloud Lupin) send audio to the company’s servers; review each provider’s data retention policy before using for sensitive content.
Can I use voice typing in any app?
Yes if Gboard is your keyboard. The microphone button appears in any text field. Some specialized apps have their own dictation; Otter and Whisper-based apps run in their own UI as standalone.
What about accents and non-native speakers?
Whisper handles accents better than Gboard, especially for non-native English speakers. Otter’s speaker identification is also strong for accented voices in mixed-language meetings. For mid-range to budget phones, Gboard’s older models were weaker on accents; the 2025 update closed most of the gap.
Is there a free Whisper app I can use offline?
Whispr is the closest, but the on-device Whisper Large v3 requires recent flagship-class hardware. On a phone with 6 GB of RAM and older, the smaller Whisper variants run offline but with reduced accuracy. Gboard offline is the better baseline for budget hardware.
How does Pixel’s recorder app fit in?
Recorder (Pixels only, preinstalled) does on-device transcription of recorded audio with speaker identification and search. It is excellent for podcast-style or interview workflows. Not available on non-Pixel Androids; install Whispr or Otter for similar functionality elsewhere.
The verdict
Speech-to-text on Android in 2026 is genuinely good. Most users do not need to install anything; Gboard Voice Typing handles the daily dictation use case with no friction. The third-party apps add value for specific cases (meetings, max accuracy, accessibility) but are not strictly necessary for general use.
For meeting transcripts, install Otter. For maximum accuracy on important work, install a Whisper-powered app. For accessibility, install Live Transcribe. Each is the best in its niche; pick by what you actually need.
Audio-first input is finally on par with typing for many use cases. Spend a week dictating instead of typing for everything that does not require precision and you will probably keep the habit.
How we put this guide together
We tested Gboard (15.4 with the 2026 Voice Typing update), Lupin (3.8), Whispr (1.4), Otter (4.1), Rev Voice Recorder (2.8), and Live Transcribe (3.7) on a Pixel 8 Pro and a Galaxy S24 in May 2026. Accuracy was measured against a controlled audio set (10 minutes of varied content: business prose, casual conversation, technical jargon, accented English) and a court-reporter-verified transcript. Offline performance was confirmed by airplane mode tests. We update this guide annually or when a major model update lands.
