Privacy-First Voice AI

Voice technology that respects your data.

In a world of data-hungry voice assistants, there is a better way. TigreGótico builds voice technology on privacy, open source, and data sovereignty. As creators of the HiveMind stack and core contributors to OpenVoiceOS, we help you ship secure, GDPR-compliant voice solutions that run on your own hardware — no data sent to anyone else’s cloud.

Creators ofHiveMind& core contributors toOpenVoiceOS

What we do

Voice technology — and the data engineering behind it

We build privacy-first voice across the whole open-source stack, and we turn hard-to-reach public data into clean APIs and datasets.

Data extraction, clean APIs & datasets

A first-tier specialty: we unlock public data that is otherwise hard to reach — behind unstructured pages, anti-bot walls, or undocumented endpoints — with resilient scrapers and reverse-engineered APIs, then re-expose it as clean, typed client libraries and build curated, ML-ready datasets from it.

Data services

Privacy-first voice assistants

On-device OpenVoiceOS deployments that keep audio and data on your own hardware — GDPR-compliant by design.

Custom voice models — TTS & ASR

Offline text-to-speech voices for your language or brand, and fine-tuned speech recognition (Zipformer, Parakeet, Whisper) for your domain — backed by our ASR partnership with Alpha Cephei.

Minority & Lusophone speech tech

Phonemization, ASR, and TTS for under-served languages that mainstream tools overlook.

HiveMind device meshes

Distributed voice across satellites and hubs with fine-grained permissions and no central cloud.

Creators ofHiveMind& core contributors toOpenVoiceOS.


The stack we build on

OpenVoiceOSHiveMind

With OpenVoiceOS

Two voices. Every language.

Miro and Dii are two voice identities that sound the same no matter which language your assistant speaks — a truly universal brand voice for OpenVoiceOS. We train a monolingual model per language with our voice-cloning pipeline, doing grapheme-to-phoneme research along the way — and we cover endangered languages that mainstream tools ignore, like Frisian, Asturian, and Aragonese. Each voice is atiny ~15.6M-parameter model that runs on a Raspberry Pi, a phone, or right in your browser — no GPU, no cloud.

Featured

What we build & publish

Everything we build is FOSS and self-hostable — no lock-in, no mandatory cloud.

Our work spans the whole voice stack — text-to-speech, speech recognition, and wake-word detection, plus the OpenVoiceOS plugins that tie them together and the HiveMind mesh that runs assistants across many devices. On the data side, a family of scraping, anti-bot transport, and entity-resolution libraries turns hard-to-reach public web data into clean, typed APIs and reproducible datasets. Each piece is a small, composable library you can run on your own hardware — and the datasets and models behind them, including our Miro and Dii multilingual voices and Portuguese language tooling, are published openly on Hugging Face.

Browse all open source Datasets & models

Play

Text & voice games

From classic interactive fiction to LLM-powered generative worlds — playable in your command line or entirely by voice, designed audio-first for everyone.

From the blog

Latest writing

Read the blog

Testimonials

What partners & users say

Words from the people and projects we build with.

“If you need privacy-first voice and AI solutions, TigreGotico is our top pick for a partner. Privacy is part of their DNA. You won't find another organization as dedicated to keeping your data local and safe.”
Mike J. GrayOscillate Labs
“Casimiro did excellent work on ATC transcription for PilotGPT. He was thoughtful, technically strong, and able to work through a difficult real-world audio problem with speed and creativity. He consistently delivered high-quality work and was great to collaborate with. I'd strongly recommend him for any AI/audio engineering role.”
Murti Hussain, Founder @ PilotGPTPilotGPT
“Casimiro has been one of the driving forces behind making privacy-first voice assistants practical instead of just aspirational. His work on HiveMind and OpenVoiceOS balances strong engineering with a deep respect for privacy, openness, and long-term maintainability. When he builds infrastructure, it's designed to empower developers and users—not to lock them into someone else's ecosystem.”
Gaëtan Trellu, Founder @ ThalovantThalovant
“Casimiro is a one-of-a-kind Python developer and AI specialist. What makes him unique is his creative ability to build custom solutions in the open-source space, paired with a healthy, critical, problem-solving attitude instead of just doing what is asked. Combine that with deep knowledge of how his products work — creating the datasets, models, and workflows himself — and you always get a state-of-the-art solution.”
Timon van Hasselt, VisioLabKoninklijke Visio

Let's build

Have a voice project in mind?

Talk to the people who build the stack. No data brokers, no lock-in — just private, open voice technology built to last.