riffado/riffado

Self-hosted AI transcription interface for Plaud Note devices

What it solves

Riffado is a self-hosted companion app for AI voice recorders, specifically designed to give users more control over their recordings and transcripts. It removes the dependency on manufacturer-controlled clouds by allowing users to sync recordings from a device's cloud and store them on their own infrastructure, using their own AI providers for transcription.

How it works

The system syncs recordings from the manufacturer's cloud (currently supporting the Plaud Note family) and processes them using an OpenAI-compatible API or directly in the browser via Transformers.js (Whisper in WebAssembly). Users can choose their own storage backend, such as a local filesystem or S3-compatible storage (AWS S3, Cloudflare R2, MinIO, etc.), and all sensitive data, including API keys and transcripts, is encrypted at rest using AES-256-GCM.

Who it’s for

It is aimed at users of AI voice recorders who want to own their data, avoid vendor lock-in, and maintain high levels of privacy and security through self-hosting.

Highlights

  • Provider Agnostic: Supports any OpenAI-compatible provider (OpenAI, Groq, OpenRouter, Together, LM Studio, Ollama, Azure).
  • Browser-Based Transcription: Offers free transcription via Transformers.js (Whisper in WebAssembly).
  • Flexible Storage: Supports local filesystem or S3-compatible storage.
  • Automated Sync: Includes scheduled auto-sync with browser and email notifications.
  • Secure: AES-256-GCM encryption at rest for tokens, API keys, and transcripts.
  • Developer Friendly: Provides an automation API with signed webhooks for integrations.
  • Zero-Config Deploy: Easy installation via Docker Compose.

Related

  • Project
  • Project
  • Project
  • Project
  • Project