Prompt at the speed of speech.
Tap the mic, talk, and watch your words land in the composer in real time. Dictation is built into every Wirebase chat — no voice vendor keys, no per-minute transcription bills.
From thought to prompt, no keyboard.
Browser-native speech recognition transcribes as you speak, straight into the composer. The transcript is ordinary, editable text — review it, fix a word, then send. You stay in control of exactly what reaches the model.
- Tap the mic in any chat — desktop or mobile browser
- Live transcription into the composer as you speak
- Review and edit the transcript before sending
- Switch between speaking and typing mid-message
Dictate to every model you use.
Dictation is an input mode of the same chat, not a separate product. Spoken prompts go to whichever model the thread is using — with the same agents, connectors, and knowledge bases as typed ones. There is nothing extra to configure.
- Works with any enabled provider — OpenAI, Anthropic, Google, xAI, and more
- Same thread, same tools, same @-mentions
- No separate voice product for admins to set up
- Falls back to typing on browsers without speech support
No voice stack to buy or run.
Speech recognition runs on the browser's built-in engine, so there are no speech-to-text vendor contracts, no per-minute fees, and no new data processor for procurement to review. Wirebase only ever receives the text — audio is never uploaded to or stored on Wirebase servers.
- Browser-native speech engine — no extra API keys
- Zero per-minute transcription costs
- Wirebase receives text only; audio is never stored on our servers
- Nothing additional to deploy on self-hosted instances
Speaking is faster than typing. Dictation turns that into faster prompts — same models, same tools, no new vendors.
Dictation that fits how you already work.
One-tap dictation
Start and stop from the composer in any chat.
Edit before sending
The transcript is editable text, never a blind send.
Saved as text
Dictated messages live in the thread like any other.
Every model
Speak to any provider your organization has enabled.
No extra keys
No speech vendor accounts or per-minute fees.
Nothing stored
Wirebase receives text only — audio is never stored.