Open-source Dicio offers privacy-focused alternative to Google’s Gemini for Android users

Dicio, a free, open-source voice assistant for Android, provides a decentralised alternative to Google’s Gemini, prioritising user privacy and local processing but with limited features compared to mainstream AI assistants.

Google is steadily steering Android users towards Gemini, but there is still a different path for those who value privacy, open-source software and tighter control over their data. One of the clearest examples is Dicio, a free voice assistant for Android that is designed to handle much of the work directly on the handset, rather than sending every request to the cloud. According to the project’s listings and repository, it is available through Google Play, F-Droid and the project’s own code base, and it supports several languages, including English, French, German, Greek, Italian, Russian and Spanish.

The software takes a more traditional route than Gemini. Instead of relying on a large language model, it first converts speech to text with Vosk, the open-source speech-recognition engine, and then matches the resulting command against known patterns. That makes it less flexible for open-ended requests, but it also reduces dependence on remote AI systems. The GitHub repository describes Dicio as a free and open-source assistant with both voice and graphical feedback, and says developers can inspect and adapt the code.

Dicio’s appeal lies in its focused set of “skills”, which cover common tasks such as timers, calls, app launching, weather checks, song lyrics, translations, searches and calculations. Store listings also say it can work with multiple input and output methods, and that individual skills can be enabled or disabled in settings. That gives users more control over what the assistant is allowed to do, even if the overall feature set remains narrower than Google’s.

Setting it up is relatively straightforward. The app needs microphone access, then a language model for Vosk must be downloaded so that speech can be processed on the device. A wake-word model is also available, allowing users to summon the assistant by voice. A F-Droid package page and a project guide both describe Dicio as installable from open-source-friendly channels, with the option of using the app as Android’s default assistant in system settings.

That said, “local” does not mean fully offline. The recognition stage can run on the phone, but some skills still need external services. Weather look-ups rely on OpenWeatherMap, while other functions may use services such as DuckDuckGo or Genius. In practice, that means Dicio reduces the amount of audio and command data sent to third parties, but it does not eliminate network dependence altogether.

The trade-off is clear. Dicio offers privacy-minded users a lightweight assistant with open-source credentials and useful day-to-day functions, but it does not match Gemini’s conversational range, system integration or broader ecosystem support. It is also less polished in some areas: reports from MakeUseOf noted that wake-word recognition and some commands were not always reliable, and that Android Auto integration is lacking. For users who mainly want hands-free timers, calls and simple device control, it is a credible alternative. For those who want a more capable AI assistant, it remains a partial substitute rather than a full replacement.

Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.