| { |
| "components": [ |
| { |
| "id": "vlm-smolvlm2-500m", |
| "type": "vlm", |
| "name": "Vision — SmolVLM2 500M", |
| "description": "Tiny, fast vision model. Sees images in chat & social, fully offline. Runs on almost any phone. Apache-2.0.", |
| "quant": "Q8_0", |
| "minRamMb": 2000, |
| "files": [ |
| { "role": "model", "url": "https://huggingface.co/ggml-org/SmolVLM2-500M-Video-Instruct-GGUF/resolve/main/SmolVLM2-500M-Video-Instruct-Q8_0.gguf", "sizeBytes": 436808704, "sha256": "6f67b8036b2469fcd71728702720c6b51aebd759b78137a8120733b4d66438bc" }, |
| { "role": "mmproj", "url": "https://huggingface.co/ggml-org/SmolVLM2-500M-Video-Instruct-GGUF/resolve/main/mmproj-SmolVLM2-500M-Video-Instruct-Q8_0.gguf", "sizeBytes": 108785184, "sha256": "921dc7e259f308e5b027111fa185efcbf33db13f6e35749ddf7f5cdb60ef520b" } |
| ] |
| }, |
| { |
| "id": "vlm-qwen2.5-vl-3b", |
| "type": "vlm", |
| "name": "Vision — Qwen2.5-VL 3B", |
| "description": "Balanced vision + chat brain. Sees images, fully offline. Needs a 6 GB+ phone. License: Qwen.", |
| "quant": "Q4_K_M", |
| "minRamMb": 6000, |
| "files": [ |
| { "role": "model", "url": "https://huggingface.co/ggml-org/Qwen2.5-VL-3B-Instruct-GGUF/resolve/main/Qwen2.5-VL-3B-Instruct-Q4_K_M.gguf", "sizeBytes": 1929901056, "sha256": "d02fe9b69ad8cadbbd228e387667af66612c44bed29ffc8eb1e7caf9ac486c12" }, |
| { "role": "mmproj", "url": "https://huggingface.co/ggml-org/Qwen2.5-VL-3B-Instruct-GGUF/resolve/main/mmproj-Qwen2.5-VL-3B-Instruct-Q8_0.gguf", "sizeBytes": 844757728, "sha256": "980c9b2f78c04e6cff93d277ada09e768394f112d75db3b4e9dea8a69f9fb904" } |
| ] |
| }, |
| { |
| "id": "vlm-gemma3-4b", |
| "type": "vlm", |
| "name": "Vision — Gemma 3 4B", |
| "description": "Highest-quality vision (slower to process images). Fully offline. Needs a 6 GB+ phone. License: Gemma.", |
| "quant": "Q4_K_M", |
| "minRamMb": 6000, |
| "files": [ |
| { "role": "model", "url": "https://huggingface.co/ggml-org/gemma-3-4b-it-GGUF/resolve/main/gemma-3-4b-it-Q4_K_M.gguf", "sizeBytes": 2489757856, "sha256": "882e8d2db44dc554fb0ea5077cb7e4bc49e7342a1f0da57901c0802ea21a0863" }, |
| { "role": "mmproj", "url": "https://huggingface.co/ggml-org/gemma-3-4b-it-GGUF/resolve/main/mmproj-model-f16.gguf", "sizeBytes": 851251104, "sha256": "8c0fb064b019a6972856aaae2c7e4792858af3ca4561be2dbf649123ba6c40cb" } |
| ] |
| }, |
| { |
| "id": "vlm-gemma4-e2b", |
| "type": "vlm", |
| "name": "Vision — Gemma 4 E2B", |
| "description": "Google's newest compact vision + chat brain. Sees images, fully offline. Larger than its name (MatFormer) — needs a 6 GB+ phone. License: Gemma.", |
| "quant": "Q4_K_M", |
| "minRamMb": 6000, |
| "files": [ |
| { "role": "model", "url": "https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF/resolve/main/gemma-4-E2B-it-Q4_K_M.gguf", "sizeBytes": 3106736256, "sha256": "9378bc471710229ef165709b62e34bfb62231420ddaf6d729e727305b5b8672d" }, |
| { "role": "mmproj", "url": "https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF/resolve/main/mmproj-F16.gguf", "sizeBytes": 985654080, "sha256": "140be8d7849741f88c50757d529b84373ee8e27052cc2236855b537f4a8215fa" } |
| ] |
| }, |
| { |
| "id": "vlm-gemma4-e4b", |
| "type": "vlm", |
| "name": "Vision — Gemma 4 E4B", |
| "description": "Google's newest high-quality vision + chat brain. Larger and slower — needs an 8 GB+ phone. License: Gemma.", |
| "quant": "Q4_K_M", |
| "minRamMb": 8000, |
| "files": [ |
| { "role": "model", "url": "https://huggingface.co/unsloth/gemma-4-E4B-it-GGUF/resolve/main/gemma-4-E4B-it-Q4_K_M.gguf", "sizeBytes": 4977169568, "sha256": "519b9793ed6ce0ff530f1b7c96e848e08e49e7af4d57bb97f76215963a54146d" }, |
| { "role": "mmproj", "url": "https://huggingface.co/unsloth/gemma-4-E4B-it-GGUF/resolve/main/mmproj-F16.gguf", "sizeBytes": 990372672, "sha256": "ddf46c21d7078e95338cfc22306b19b276a29a5ad089023449dd54d4b6170a51" } |
| ] |
| }, |
| { |
| "id": "chat-llama3.2-3b", |
| "type": "chat", |
| "name": "Chat — Llama 3.2 3B", |
| "description": "Well-rounded text-only chat brain. Needs a 6 GB+ phone. License: Llama 3.2.", |
| "quant": "Q4_K_M", |
| "minRamMb": 6000, |
| "url": "https://huggingface.co/bartowski/Llama-3.2-3B-Instruct-GGUF/resolve/main/Llama-3.2-3B-Instruct-Q4_K_M.gguf", |
| "sizeBytes": 2019377696, |
| "sha256": "6c1a2b41161032677be168d354123594c0e6e67d2b9227c84f296ad037c728ff" |
| }, |
| { |
| "id": "root-qwen2.5-1.5b", |
| "type": "agent", |
| "name": "Root Agent — Qwen2.5 1.5B", |
| "description": "Recommended Root brain for background jobs (memory, dossier, App Guide). Small & fast. Apache-2.0.", |
| "quant": "Q4_K_M", |
| "minRamMb": 3000, |
| "url": "https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct-GGUF/resolve/main/qwen2.5-1.5b-instruct-q4_k_m.gguf", |
| "sizeBytes": 1117320736, |
| "sha256": "6a1a2eb6d15622bf3c96857206351ba97e1af16c30d7a74ee38970e434e9407e" |
| }, |
| { |
| "id": "embed-minilm-l6-v2", |
| "type": "embed", |
| "name": "Semantic Memory — all-MiniLM-L6-v2", |
| "description": "Tiny embedding model that powers semantic (hybrid) memory recall — characters surface relevant memories even when your wording differs. Fully offline, runs on any phone. Apache-2.0.", |
| "quant": "F16", |
| "minRamMb": 1000, |
| "url": "https://huggingface.co/second-state/All-MiniLM-L6-v2-Embedding-GGUF/resolve/main/all-MiniLM-L6-v2-ggml-model-f16.gguf", |
| "sizeBytes": 45949216, |
| "sha256": "797b70c4edf85907fe0a49eb85811256f65fa0f7bf52166b147fd16be2be4662" |
| }, |
| { |
| "id": "asr-whisper-base", |
| "type": "asr", |
| "name": "Speech-to-Text — Whisper Base", |
| "description": "On-device voice transcription for calls — fully offline, multilingual. Recommended balance of accuracy and speed. Runs on any phone. MIT.", |
| "quant": "ggml", |
| "minRamMb": 1000, |
| "url": "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.bin", |
| "sizeBytes": 147951465, |
| "sha256": "60ed5bc3dd14eea856493d334349b405782ddcaf0028d4b5df4088345fba2efe" |
| }, |
| { |
| "id": "asr-whisper-tiny", |
| "type": "asr", |
| "name": "Speech-to-Text — Whisper Tiny", |
| "description": "Smallest on-device voice transcription for calls — fully offline, multilingual. Fastest, slightly lower accuracy. Runs on any phone. MIT.", |
| "quant": "ggml", |
| "minRamMb": 1000, |
| "url": "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-tiny.bin", |
| "sizeBytes": 77691713, |
| "sha256": "be07e048e1e599ad46341c8d2a135645097a538221678b7acdd1b1919c6e1b21" |
| }, |
| { |
| "id": "asr-whisper-large-v3-turbo", |
| "type": "asr", |
| "name": "Speech-to-Text — Whisper Turbo", |
| "description": "Most accurate on-device voice transcription for calls — fully offline, multilingual. Larger download; best on a 6 GB+ phone. MIT.", |
| "quant": "ggml q5_0", |
| "minRamMb": 3000, |
| "url": "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-large-v3-turbo-q5_0.bin", |
| "sizeBytes": 574041195, |
| "sha256": "394221709cd5ad1f40c46e6031ca61bce88931e6e088c188294c6d5a55ffa7e2" |
| }, |
| { |
| "id": "tts-en-amy", |
| "type": "tts", |
| "name": "Voice — English (Amy)", |
| "description": "Lets your characters speak their replies aloud in calls — fully offline, no API key needed. Runs on any phone.", |
| "minRamMb": 1000, |
| "url": "https://huggingface.co/Mr-J-369/Fancy-AI/resolve/main/tts-en-amy-low.zip", |
| "sizeBytes": 67308190, |
| "sha256": "ec2dafc9b8efa91a033a083a0ab50be1c9f1715e3d33e8ab22e4d969a553dfd3" |
| } |
| ] |
| } |
