Do work with voice.
Freestyle turns your speech into polished text, helping you write faster and get work done.
Speak naturally. It writes perfectly.
Hold the hotkey, talk, then release. Polished text lands at your cursor in milliseconds. Freestyle can write into any app with a text field. Support for 80 languages.
GmailSlackNotionDocsWhatsAppNotesFigmaTelegram
macOS
Windows
Linux
iPhoneComing soon
AndroidComing soon
Freestyle edits while you speak.
Freestyle cleans out filler words, corrects punctuation, and formats what you say into clean text.
Fillers, gone
“Ahs” and “uhms” never reach the page.
so um can we uh move the standup to ten
Can we move the standup to 10?
Punctuation that listens
Inferred from your pauses and tone.
she said whats the timeline
She said, “What’s the timeline?”
Backtrack
Change your mind mid-sentence; only the correction lands.
meet at two… actually three
Let’s meet at 3.
Spoken lists
Numbers become formatted lists.
one venue two budget three agenda
1. Venue
2. Budget
3. Agenda
2. Budget
3. Agenda
Sounds like you.
Teach Freestyle your world once. It spells your teammates, your products, and your industry right, everywhere.
Dictionarylearns as you go
Syncing the Kubernetes cluster after standup.
Saved to your dictionary. Spelled right everywhere, forever.
Next time: “ …deploying to Kubernetes tonight.” ✓
Vocabularyyour industry, understood
Load your world in once, and technical dictation lands right the first time.
lisinoprilSeries BSupabasecamelCaseMRRrefactor
recognized live while you dictate
Tones
Formal in docs, casual in chat, warm in email. Freestyle shifts register based on where you're writing.
Formal · DocsCasual · SlackWarm · Email
Build your own dictation experience.
Flip on Dev Mode and choose what powers your transcription. Run models on your own machine, or bring your own keys.
Local models
Transcription runs entirely on your machine. No account, no network, no per-minute cost.
Qwen3 ASR
0.6B, 8-bit
SenseVoice
multilingual
Parakeet
NVIDIA NeMo
w
whisper.cpp
tiny to large
Cloud providers
Plug in your keys and pay providers directly. Swap models at any step of the pipeline.
OpenAI
gpt-4o
G
Groq
whisper turbo
Deepgram
nova-3
ElevenLabs
scribe
Private by default.
We never store your transcriptions, or any data of any kind, on free and paid plans alike. Privacy isn't an upgrade. It's the default.
SOC 2 Type 2In progress
HIPAAIn progress
ZDRZero data retention
Privacy on every plan
Privacy isn't gated behind enterprise. Chart notes, privileged text, and secrets are safe to speak on the free tier.
Self-hosted & isolated
Source-available. Run Freestyle yourself, or let us host it in an isolated container, with a dictation experience built for your team.