Free web
Free source (opens in a new tab)
One-time
For: Anyone dictating on a Mac
Limits
- macOS 15.0 or later
- Apple Silicon and Intel
Includes
- System-wide hotkey dictation
- On-device speech models
1 more
- Optional Fluid-1 polish
No API
Audio & voiceDocumented overviewLast checked Sep 29, 2026
FluidVoice: Pricing, Features & Best Use Cases
Free open-source macOS dictation app that transcribes speech on-device and polishes it with an optional local model.
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)
Our take: FluidVoice is the pick for Mac users who want free, open-source dictation into any app with transcription kept on the device and an optional local model for cleanup.
Main limitation: It needs macOS 15.0 or later, microphone and accessibility permissions, and the optional Fluid-1 model is a roughly 3.5 GB download.
Who should use FluidVoice
FluidVoice is the pick for Mac users who want free, open-source dictation into any app with transcription kept on the device and an optional local model for cleanup. Based on the vendor's documentation, not on a hands-on test.
Best for: Mac users who dictate often
Also a good fit for
- Developers
- Privacy-focused users
Consider another tool if macOS 15.0 or later required
Or if
- Optional Fluid-1 model is a ~3.5 GB download
- Language support varies by model
FluidVoice in the news
Headlines from US outlets, selected automatically — not endorsements. Updated Oct 2, 2026
FluidVoice pricing and usage limits
Prices checked Sep 29, 2026 on the official US pricing page. Official pricing ↗ Prices and limits change often.
Free access and trial conditions
Free forever, open source (GPLv3), no tiers
Not applicable
Paid plans: what changes as you upgrade
Usage caps, credits and extra costs
Optional cloud providers (OpenAI, Groq) bill separately if you use your own keys
Not applicable
What is FluidVoice, and how does it work?
FluidVoice is a macOS dictation app that types transcribed speech into any app via a hotkey. Transcription runs locally using models such as Parakeet and Whisper. An optional on-device model (Fluid-1) can clean up formatting and adapt tone per app; cloud providers like OpenAI or Groq can optionally be used instead. The site describes it as free and open source (GPLv3).
What you provide and what you get back
You provide
- Speech from the microphone via hotkey
- Custom per-app prompts
- Optional cloud provider keys
You get back
- Transcribed text typed into the active app
- Polished text with filler removed and tone matched
- Command-mode actions
Where you can use it
- Platforms
- Desktop
- Account
- Not required
- OS
- macOS 15.0+
- Speech models
- Parakeet, Whisper, Nemotron, Apple Speech
- Languages
- 40+ across models
- Modes
- Write, Command, Dictation
- License
- GPLv3
What people use FluidVoice for
Scenarios are worked out by our editors from the documentation; the actual result depends on your input and plan.
Dictate emails and chat messages
- Who
- Mac users who write a lot
- You provide
- Spoken sentences with a hotkey
- Expected result
- Clean text typed into Mail or Slack
- Plan needed
- Free
- Limitation
- Language quality varies by model source (opens in a new tab)
Dictate prompts into coding tools
- Who
- Developers using terminals and AI assistants
- You provide
- Rambling spoken instructions
- Expected result
- A tidy prompt in Claude Code or similar
- Plan needed
- Free
- Limitation
- Needs accessibility permission to type source (opens in a new tab)
Offline transcription
- Who
- Privacy-minded users
- You provide
- Audio with no network
- Expected result
- Text produced fully on-device
- Plan needed
- Free
- Limitation
- Cloud providers are optional, not required source (opens in a new tab)
Per-app tone control
- Who
- Users with different writing styles
- You provide
- Custom prompts for Slack, Mail or GitHub
- Expected result
- Casual, formal or structured output per app
- Plan needed
- Free
- Limitation
- You write and maintain the prompts source (opens in a new tab)
What FluidVoice does, and where it stops
System-wide hotkey dictation
Types transcribed speech into any app when you hold a hotkey No copy and paste between tools
On-device speech models
Runs Parakeet, Whisper, Nemotron, Cohere Transcribe and Apple Speech locally Keeps audio off servers and works offline
Optional Fluid-1 local AI polish
Removes filler, fixes casing and matches tone using a local model Turns rough speech into send-ready text
Write, Command and Dictation modes
Switches between writing, commanding and direct dictation Fits different tasks without changing apps
Cloud provider option
Lets you use OpenAI, Groq or custom providers instead of local models Offers a choice of speed and accuracy
FluidVoice pros and cons
Reasons to choose FluidVoice
- Free, no tiers, per the site Nothing to pay for dictation documented source (opens in a new tab)
- Open source (GPLv3) Code can be audited and forked documented source (opens in a new tab)
- Local-first processing Voice stays on your Mac by default documented source (opens in a new tab)
- Intel and Apple Silicon supported Works on a wide range of Macs documented source (opens in a new tab)
- 40+ languages across models Multilingual dictation is possible documented source (opens in a new tab)
Check before paying for FluidVoice
- macOS 15.0 or later required Older Macs are excluded documented source (opens in a new tab)
- Optional Fluid-1 model is a ~3.5 GB download Needs disk space and a long first download documented source (opens in a new tab)
- Language support varies by model Some languages work worse than others documented source (opens in a new tab)
- Needs microphone and accessibility permissions Requires system settings changes before use documented source (opens in a new tab)
- No Windows or iOS release Only a beta teaser exists for other platforms documented source (opens in a new tab)
How to use FluidVoice for dictating text into any Mac app
Set up FluidVoice
- Install with brew install --cask fluidvoice or download from altic.dev/fluid
- Grant microphone and accessibility permissions in System Settings
- Choose a speech model such as Parakeet or Whisper and let it download
- Optionally download Fluid-1 for polish
Example: dictating text into any Mac app
- Click into any text field
- Hold the hotkey and speak your message
- Release and watch the text appear
- Switch to Write mode for a polished version
Review and export the result
- Confirm the text lands in the right app
- Disconnect Wi-Fi and verify dictation still works
FluidVoice API
API access and separate billing
FluidVoice has no public API in its official documentation. Local app; cloud providers use your own keys API documentation ↗
Privacy, data use and commercial rights
Summarized from the vendor's current policies; this is not legal advice and a vendor's badge is not an independent audit.
How your inputs and outputs are handled
Audio is transcribed on-device; optional cloud providers receive data only if you enable them
Site says zero data leaves your Mac by default
Not stated
Commercial use and plan restrictions
GPLv3 open source
None
Policies: altic.dev/fluid/privacy
FluidVoice alternatives by need
"Better suited" is based on documented features and pricing, not on a head-to-head test.
Otter.ai: better suited to
Explore FluidVoice further
Related to FluidVoice
More audio & voice tools
Guides and comparisons
FluidVoice user reviews
No user reviews yet.
Used FluidVoice? Be the first to share your experience.
Write a review
FluidVoice FAQ
Is FluidVoice really free?
Yes, the site says it is free forever and open source under GPLv3, with no tiers. source (opens in a new tab)
Does my voice leave my Mac?
Not by default: transcription runs on-device, and cloud providers are used only if you choose them. source (opens in a new tab)
Which macOS versions work?
macOS 15.0 or later on Apple Silicon and Intel. source (opens in a new tab)
What is Fluid-1?
A custom-trained local model that cleans up and formats transcripts, about a 3.5 GB download. source (opens in a new tab)
Which speech models are included?
Nemotron Speech, Parakeet variants, Cohere Transcribe, Apple Speech and Whisper sizes. source (opens in a new tab)
Sources and review methodology
Official sources and verification dates
- FluidVoice product page homepage · checked Oct 1, 2026
- Altic home page (publisher) homepage · checked Oct 1, 2026
Not verified: Per-model language list; Retention policy details; US availability.
How we documented and maintain this page
Every fact about FluidVoice above comes from the 2 official pages listed here, last compared on Sep 29, 2026; we have not run FluidVoice ourselves, so it has no editorial score. How we check and re-check pages.
Article updated Oct 10, 2026; prices checked Sep 29, 2026. Outbound links to vendors are plain links; we do not earn commissions. How we review · Report an error