Direct answer — Is VoiceGecko a good free dictation app for Windows? Yes, if Windows is your only platform and you want something free and auditable. VoiceGecko is MIT-licensed, runs Whisper models locally, keeps transcripts in a local SQLite database and needs no account to dictate. The trade-offs are a Windows-only release, a young project with a small community, no published language list, and support limited to GitHub and Discord.
This VoiceGecko review looks at a newer entrant among free, local dictation tools for Windows. VoiceGecko presents itself as an open-source alternative to Wispr Flow: press a shortcut, speak, and get your words as text without sending audio to a cloud service.
We checked every claim below against VoiceGecko’s website, its GitHub README, its release history and parts of its source code. Where something is not documented, we say so.
What is VoiceGecko?
VoiceGecko is a free, open-source desktop dictation app, currently released for Windows only. Its GitHub description calls it “open source, local voice-to-text (Wispr Flow alternative)”, and its website promises “instant voice dictation at your fingertips”.
The README describes the workflow in one line: “Press a shortcut, speak, get text on your clipboard.” From there you paste into Slack, VS Code, Jira, a browser or any other desktop app. The website also mentions the GeckoBar, a small on-screen bar that can start a dictation.
Some background facts, all from the repository:
- Publisher: the licence names Social Freak Limited, trading as Voice Gecko
- Technology: a Tauri desktop app, with a separate website built on Next.js
- Latest release: v2.0.3, published on 15 September 2026
- Community: 35 stars on GitHub at the time of writing, which makes it a small, early-stage project
Is VoiceGecko really free and open source?
Yes. The code is published under the MIT licence, which lets you read, modify, fork and reuse it, including in commercial projects. The pricing page lists a Free plan described as “Full product. No limits.”
There is also an optional paid Support subscription. VoiceGecko states that “Free and Support are the same product — no paywalled features”: paying funds development, it does not unlock anything. An account is only needed for Support checkout or to send feedback, not to dictate.
That is an unusually clean model. The flip side is the same as for any community project: support runs through GitHub issues and a Discord server, with no service commitment.
Does VoiceGecko work offline, and what stays on your PC?
Yes, for transcription. VoiceGecko says that “speech runs 100% locally on your device” and that audio is “processed locally and never uploaded”. According to the README, speech models download once on first launch, and dictation then works offline.
| Data | Where it goes, according to VoiceGecko |
|---|---|
| Your audio | Processed on your PC, never uploaded |
| Transcripts | Stored in a local SQLite database, not synced |
| Custom dictionary | Stored locally with the rest of your data |
| Speech models | Downloaded once on first launch |
| Usage analytics | PostHog events in the desktop app, behind a Usage Analytics setting |
The last row deserves a note. The desktop source code includes a PostHog analytics module, and its documentation says tracking is skipped when you turn off Usage Analytics in the privacy settings. The website does not mention this, so check that setting if you want nothing but model downloads leaving your machine.
Which speech engine and languages does VoiceGecko use?
VoiceGecko runs OpenAI Whisper locally. Its README mentions “local speech models (Parakeet / Moonshine / Whisper)”, but the desktop code we read manages Whisper models only.
Under the hood, a separate process runs Whisper through Whisper.net, a .NET wrapper that ships native whisper.cpp builds. The model manager offers Whisper sizes from Tiny to Large V3, including quantised versions and English-only variants. The default is small.en, an English-only model.
On languages, VoiceGecko publishes no list. The multilingual Whisper models cover many languages, but the Whisper project itself warns that performance “varies widely depending on the language”. If you dictate in anything other than English, switch away from the default model and test your own language first.
The code also includes an optional post-dictation polish step that uses a local language model, Qwen2.5 3B, downloaded (about 2 GB) only when you enable it.
What features does VoiceGecko include?
VoiceGecko keeps the feature list short. Based on its website, pricing page and README, you get:
- Global shortcuts, customisable, to start and stop a dictation
- A custom dictionary for names, product terms and industry jargon
- Unlimited dictations, with no weekly or word limits
- Local transcript history in SQLite
- The GeckoBar, an on-screen bar for quick access
- Optional local polish of the dictated text
The website also states that people speak at about 200 words per minute and type at about 40, and headlines VoiceGecko as “5x faster than your keyboard”. It does not say how, or on whom, these figures were measured, so treat them as marketing rather than benchmarks.
What are VoiceGecko’s limits?
The main limit is platform: VoiceGecko is Windows-only. The v2.0.3 release ships a single Windows x64 MSI installer. The download page lists macOS, Linux, web, Android and iOS as future platforms that visitors can vote for, with no dates.
Other points to weigh:
- No published system requirements: neither the README nor the download page text lists a minimum Windows version, CPU or RAM.
- No ARM64 build: only an x64 installer appears in the latest release.
- Clipboard output: the README describes text arriving on your clipboard, so expect a paste step depending on your workflow.
- Young project: with 35 GitHub stars and a small contributor base, long-term maintenance depends on a small team.
None of these makes VoiceGecko a bad choice. They mean you should test it on your own hardware before relying on it for daily work. If you are still choosing a setup, our Windows 11 voice dictation guide covers the built-in options first.
VoiceGecko vs Weesper: how do they compare?
Both apps run Whisper locally and start dictation from a global shortcut. The difference is the kind of product: VoiceGecko is a free community project for Windows; Weesper Neon Flow is a commercial app for Mac and Windows with vendor support and no cloud features.
| Weesper Neon Flow | VoiceGecko | |
|---|---|---|
| Platforms | macOS 13+, Windows 10 (19041+) / 11 | Windows (x64 MSI) |
| Licence | Commercial | MIT open source |
| Local engine | Whisper | Whisper via Whisper.net |
| Cloud features | None | None for dictation; usage analytics behind a setting |
| Languages (vendor claims) | 55+ | Not documented |
| Custom vocabulary | Yes | Yes (custom dictionary) |
| Support | Vendor support and Help Center | GitHub issues and Discord |
| Price | 15-day trial, then €5/mo, €45/yr or €99 lifetime | Free (optional Support subscription) |
Where VoiceGecko is stronger: it costs nothing, its code is auditable under MIT, and it puts no feature behind a paywall. For a developer who only uses Windows and wants to read the code before trusting it, that is a strong offer.
Where Weesper Neon Flow is stronger: it runs on both macOS and Windows, documents 55+ languages, and comes with vendor support and a Help Center when something breaks. If you want to try it on your own documents, start the 15-day Weesper trial, no credit card required.
Other free, local options are worth a look too. Our Handy review covers a cross-platform MIT alternative, our WhisperStream review looks at another Windows offline tool, and the Weesper comparison hub lists head-to-head pages.
Free open-source tool or paid offline app: what is the real trade-off?
The trade-off is maintenance, not privacy: both keep your audio on your device. With VoiceGecko, you choose the model, check the analytics setting and troubleshoot through GitHub. With a paid tool, you pay to hand that work to a vendor.
For Weesper Neon Flow’s own plans, the cost over time is:
| Weesper Neon Flow plan | After 3 years | After 5 years |
|---|---|---|
| Monthly (€5/month) | €5 × 36 = €180 | €5 × 60 = €300 |
| Annual (€45/year) | €45 × 3 = €135 | €45 × 5 = €225 |
| Lifetime (€99 once) | €99 | €99 |
Who should choose VoiceGecko, and who should look elsewhere?
Choose VoiceGecko if you use Windows only, dictate mostly in English, and want a free tool whose code you can inspect. A developer dictating commit messages, chat replies and tickets on a Windows desktop, for example, gets local dictation at no cost.
Look for a VoiceGecko alternative if you also work on a Mac, dictate in several languages, or need a vendor to answer when dictation fails before a deadline. A consultant switching between a Windows laptop and a MacBook, for instance, needs the same tool on both.
If that sounds like you, try Weesper Neon Flow free for 15 days and dictate your own work in both apps side by side.
Verdict
In our assessment, VoiceGecko is a credible free option for local dictation on Windows. Its privacy promise is backed by public MIT code, its pricing puts no feature behind a paywall, and it uses the proven Whisper engine.
Its weaknesses are those of an early community project: Windows only, no published language list or system requirements, and an English-only default model. For English dictation on a single Windows PC, those are easy trades. For multilingual or cross-platform professional work, a supported offline tool removes the guesswork.