Voice input
that feels native
Add state-of-the-art dictation to your app or website with one SDK, so users can write by speaking
- Web
- iOS
- Android
iOS
Android Cloudflare protection... 00:00
People have more to say than they want to type
2.5x
Longer answers
+48%
More specific queries
3 in 4
Users prefer speaking
3x
Faster input
Based on research from
Cross-platform
One voice layer across Web, iOS, and Android.
- One SDK
- Shared events
- Every surface
Fast
Streaming transcription keeps every thought moving.
- Live streaming
- Instant feedback
- Low-latency insert
Accurate
Precise words and punctuation, right where users need them.
- Clean transcripts
- Smart punctuation
- Stable final text
Satisfying
A fluid voice experience users enjoy coming back to.
- Clear states
- Smooth recovery
- Natural finish
May 19, 2026Voice speeds up writing
“In “voice mode”, you can produce double the words-per-minute than you can when typing.”
Dec 1, 2025Voice builds trust
“Our research found that trust increases 33% when people speak to AI instead of typing.”
Jul 7, 2026Voice replaces typing
“I’m already looking for more and more opportunities to use voice instead of having to type.”
Jul 27, 2025Voice removes friction
“The friction of formatting, of reducing ideas into clickable, typable commands, disappears.”
Cross-platform
One voice layer across Web, iOS, and Android.
- One SDK
- Shared events
- Every surface
Fast
Streaming transcription keeps every thought moving.
- Live streaming
- Instant feedback
- Low-latency insert
Accurate
Precise words and punctuation, right where users need them.
- Clean transcripts
- Smart punctuation
- Stable final text
Satisfying
A fluid voice experience users enjoy coming back to.
- Clear states
- Smooth recovery
- Natural finish
May 19, 2026Voice speeds up writing
“In “voice mode”, you can produce double the words-per-minute than you can when typing.”
Dec 1, 2025Voice builds trust
“Our research found that trust increases 33% when people speak to AI instead of typing.”
Jul 7, 2026Voice replaces typing
“I’m already looking for more and more opportunities to use voice instead of having to type.”
Jul 27, 2025Voice removes friction
“The friction of formatting, of reducing ideas into clickable, typable commands, disappears.”
Design inspiration
Cloudflare protection...
Use the skill.md to build first-class voice UI
State
A state shows the current dictation phase. For example:
- idle
- listening
- transcribing
- failed
Events
An event tells your UI when dictation changes. For example:
- state_changed
- volume_changed
- transcript_delta
- text_inserted
Route every request through frontier models
Stream audio both ways while the session stays open.
Return clean transcripts ready to insert into your product.
Switch models on failure without interrupting the user.
- Pricing. Hourly equivalent; second-based when available, token-estimated otherwise.
- Accuracy. 100% − WER from the Artificial Analysis.
- Latency. Dictum real-world transcript return time; non-real-time measurements use P95.
Free
BRING YOUR OWN KEY
BYOK account
Dictum does not charge for BYOK usage. Your selected provider may bill you under its own pricing and terms.
- $0 service fee from Dictum
- No Dictum usage tiers or overages
- Use a compatible provider account
- Provider charges are paid directly by you
Get started
No payment card for Dictum.
FULLY MANAGED
per month
500 hours
- $39 including up to 500 hours
- $79 including up to 2,000 hours
- $149 including up to 5,000 hours
- $0.05 per additional hour beyond 5,000
Choose managed
No card required.
Adding dictation is a product upgrade
Product Value
Voice gives users a faster, more natural way to express complete ideas.
- Write more with less effort
- Capture richer, more specific responses
