Skip to content

Model catalog

Whisper catalog
CapabilityStatusCurrent behavior
Whisper TinyAvailableSmallest Whisper model exposed by the in-app transcription catalog.
Whisper BaseAvailableLightweight Whisper model exposed by the in-app transcription catalog.
Whisper SmallAvailableWhisper model recommended by the current heuristic for many lower-memory machines.
Whisper MediumAvailableWhisper model used by the current recommendation heuristic for Intel machines with sufficient memory.
Whisper Large v3 TurboAvailableDefault transcription model and the current Apple Silicon recommendation for many machines with at least 16 GB of memory.
Whisper Large v3AvailableLargest Whisper model exposed by the in-app transcription catalog.
Local and provider-aware model catalog
CapabilityStatusCurrent behavior
Llama 3.2 3BAvailableLightweight local model exposed by the curated Ollama catalog and recommendation heuristic.
Llama 3.1 8BAvailableDefault local summary and chat model and the balanced recommendation for higher-memory machines.
Llama 3.1 70BAvailableLarge optional local model exposed by the curated Ollama catalog.
Mistral 7BAvailableOptional local model exposed by the curated Ollama catalog.
Mistral NeMo 12BAvailableOptional local model exposed by the curated Ollama catalog.
Gemma 2 9BAvailableOptional local model exposed by the curated Ollama catalog.
Phi 3.5 3.8BAvailableLightweight local model exposed by the curated catalog and lower-memory recommendation path.
Qwen 2.5 7BAvailableOptional local model exposed by the curated Ollama catalog.
Nomic Embed TextBetaLocal embedding model used by the Ollama-backed search and indexing paths.
OpenAI-compatible embedding modelBetaThe configured remote endpoint must implement this embedding model and can receive transcript, summary, or query text.

Sizes in the product facts are catalog estimates, not a disk-space, network, performance, or quality guarantee. A remote endpoint must support the fixed embedding model identifier used by remote embedding paths.