What AI models does Hyprnote use?
Hyprnote uses two types of AI models: speech-to-text (STT) models for transcription and large language models (LLMs) for generating summaries and notes.
Speech-to-Text Models
Hyprnote supports multiple cloud-based STT providers for transcription. See the cloud providers section below for details.
Large Language Models
Hyprnote integrates with external LLM providers like LM Studio and Ollama for AI features like summaries and note generation. This gives you flexibility to choose the model that best fits your needs.
Recommended models:
Gemma - Google's open-source language model family, available in various sizes. Great balance of quality and performance.
Qwen - Alibaba's open-source language model with strong multilingual capabilities.
GPT-OSS - Community-driven open-source models compatible with OpenAI's API format.
See our Local LLM Setup guide for instructions on configuring LM Studio or Ollama with Hyprnote.
Does my audio data leave my device?
When using cloud-based transcription, your audio is sent to the selected STT provider for processing. See the cloud providers section below for details on each provider's data practices.
Your recordings are stored locally on your device. Transcripts and notes remain on your computer unless you explicitly choose to use cloud AI features.
What cloud transcription providers does Hyprnote support?
When cloud transcription is enabled, Hyprnote can use the following providers:
Deepgram - Real-time speech recognition with support for 30+ languages. Privacy Policy
AssemblyAI - Speech-to-text with speaker diarization. Privacy Policy
Soniox - High-accuracy transcription with support for 70+ languages. Privacy Policy
Fireworks AI - Fast inference for Whisper models. Privacy Policy
Gladia - Real-time transcription service. Privacy Policy
OpenAI - Whisper API for transcription. Privacy Policy
How is my data handled by cloud providers?
When using cloud transcription, your audio is sent to the selected provider for processing. Each provider has different data retention and privacy policies:
Most providers process audio in real-time and do not retain your data after transcription is complete. However, we recommend reviewing each provider's privacy policy for specific details about data handling, retention periods, and security measures.
Hyprnote Pro users who use cloud transcription through our proxy benefit from our agreements with these providers, which include data processing terms designed to protect your privacy.
Can I use Hyprnote completely offline?
Hyprnote can record audio offline, but transcription and AI features require an internet connection when using cloud providers. If you run a local LLM server (via LM Studio or Ollama), you can generate summaries offline after transcription is complete.
Does Hyprnote train AI models on my data?
No. Hyprnote does not use your recordings, transcripts, or notes to train AI models. When using cloud providers, your data is processed according to their respective privacy policies, but Hyprnote itself does not collect or use your data for training purposes.
How do I ensure maximum privacy?
For maximum privacy:
- Use a local LLM server (LM Studio or Ollama) for AI features instead of cloud providers
- Keep your recordings stored locally on your device
- Review the privacy policies of your chosen STT provider
- Review your settings in Settings > Intelligence
When using local LLM servers, your notes and summaries are generated on your device without being sent to external servers.