The model registry
Find your model.
Text, voice, and vision. Start with the curated registry,
or bring your own ONNX or GGUF model.
- Vision + language 27.3Bparameters 3.5 GBgguf
Bonsai 27B
PrismMLPrismML Bonsai 27B 1-bit vision-language model for local reasoning.
- Text generation 270Mparameters 278.1 MBgguf
Functiongemma 270M It
GoogleGoogle FunctionGemma 270M instruction-tuned function-calling model for local inference.
- Text generation 1Bparameters 768.7 MBgguf
Gemma 3 1B
GoogleGemma 3 1B - Google's mobile-optimized instruction-tuned LLM with 32K context
- Text generation 5.1Bparameters 2.9 GBgguf
Gemma 4 E2b
GoogleGemma 4 E2B - Google's compact multimodal LLM with 2.3B effective params, 128K context, and audio/image/video understanding
- Text generation 8Bparameters 4.6 GBgguf
Gemma 4 E4b
GoogleGemma 4 E4B - Google's mid-range multimodal LLM with 4.5B effective params, 128K context, and audio/image/video understanding
- Text generation 1Bparameters 768.7 MBgguf
Gemma3npc 1B
GoogleGemma3NPC 1B - Mobile-friendly NPC roleplay fine-tune of Gemma 3 1B for in-character game dialogue
- Text generation 7Bparameters 3.9 GBgguf
Gemma3npc It
GoogleGemma3NPC-it - NPC roleplay fine-tune of Gemma3n-E4B for in-character game dialogue
- Text to speech 40Mparameters 90.2 MBonnx
Kitten Tts Micro 0.8
KittenMLKittenTTS Micro 0.8 - Compact StyleTTS 2 (40M params, 8 named voices, OpenPhonemizer)
- Text to speech 80Mparameters 111.1 MBonnx
Kitten Tts Mini 0.8
KittenMLKittenTTS Mini 0.8 - High-quality StyleTTS 2 (80M params, 8 named voices, OpenPhonemizer)
- Text to speech 15Mparameters 18.3 MBonnx
Kitten Tts Nano 0.2
KittenMLKittenTTS Nano 0.2 - Ultra-lightweight TTS (15M params, <25MB)
- Text to speech 15Mparameters 78.3 MBonnx
Kitten Tts Nano 0.8
KittenMLKittenTTS Nano 0.8 - Ultra-lightweight StyleTTS 2 (15M params, 8 named voices, OpenPhonemizer)
- Text to speech 82Mparameters 174.9 MBonnx
Kokoro 82M
HexgradKokoro 82M - High-quality TTS with 24 voices, Misaki dictionary
Sizes show the default model file; some variants need extra assets.
Beyond the registry
Bring your own model.
Use any ONNX or GGUF model compatible with your runtime. Describe its files and execution pipeline in model_metadata.json, check device resources, then run it locally.