A self-hosted, OpenAI-compatible transcription endpoint running on dedicated GPU hardware — built and operated by Hoelee, full-stack developer & DevOps engineer.
This is a private speech-to-text server, not a public service. It exposes the standard OpenAI transcription contract so that any compatible client — a desktop dictation app, a mobile keyboard, or internal tooling — can use it. Access is keyed: requests without a valid API key are rejected.
Audio is processed on hardware we control and is not retained after transcription. There is no third-party cloud in the path.
POST /v1/audio/transcriptions
Authorization: Bearer <api-key>
→ {"text": "transcribed audio"}/v1/audio/transcriptionsKeys are issued per device and provisioned manually, so this endpoint stays private by default. Message me if you want access for your own devices or team.
I build this kind of internal infrastructure: self-hosted services with proper authentication, GPU workloads that share hardware sensibly, and the integration layer that makes them usable from the devices your staff already carry.
Speech-to-text is one example. The same approach applies to private document search, internal automation, VPN/remote access, monitoring, and backups — running on hardware you own, so your data never leaves the building.
Available for freelance and contract work — remote, worldwide.
I wrote up how this server is built — the hardware, the auth layer, and what running it actually costs: Your Own Speech-to-Text Server: Faster Than Typing, Private by Default.