InnerVault

Library

85,762 resources from 11 public sources. 54 checked by hand; the rest are unreviewed.

Results

  1. BigMoeOnEdge

    AI repositories

    Run MoE models bigger than your RAM. Frontier-size MoE on a phone, laptop, edge. Lossless, on stock llama.cpp. An adaptive streaming optimizer. Any hardware, any OS, any model, any quantization, behind a local OpenAI-compatible server.

    Indexed GitHub, Helldez 924 ★Apache-2.0Updated yesterday

    git clone https://github.com/Helldez/BigMoeOnEdge.git
  2. text generation model by whittleOrg · transformers

    Indexed Hugging Face Hub, whittleOrg 58 ♥apache-2.0Updated 34 days ago

    whittleOrg/Whittle-Next-27B-A3B
  3. text to speech model by inclusionAI · transformers

    Stale Hugging Face Hub, inclusionAI 65 ♥mitUpdated 26 days ago

    inclusionAI/Ming-3.0-audio-flash
  4. halogen-flash-server

    AI repositories

    Local inference that behaves like a frontier API: Qwen3.8-Flash-Next on AMD Strix Halo plus small NPU models, over the OpenAI and Anthropic APIs on one port.

    Stale GitHub, peonist-ai 1,258 ★Custom licence (see repository)Updated 9 days ago

    git clone https://github.com/peonist-ai/halogen-flash-server.git
  5. text to speech model by inclusionAI · transformers

    Stale Hugging Face Hub, inclusionAI 44 ♥mitUpdated 21 days ago

    inclusionAI/Ming-3.0-tts-tiny
  6. image text to text model by SC117 · gguf

    Stale Hugging Face Hub, SC117 150 ♥otherUpdated 24 days ago

    SC117/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-abliterated-GGUF
  7. text generation model by translated · transformers

    Indexed Hugging Face Hub, translated 27 ♥cc-by-nc-sa-4.0Updated 4 days ago 1 note

    translated/Lara-4-Mini-Base
  8. text to speech model by inclusionAI · transformers

    Stale Hugging Face Hub, inclusionAI 50 ♥mitUpdated 27 days ago

    inclusionAI/Ming-3.0-audio-tiny
  9. image text to text model by Agnes-AI · transformers

    Indexed Hugging Face Hub, Agnes-AI 22 ♥apache-2.0Updated 2 days ago

    Agnes-AI/Agnes-2.5-Pro-Theta
  10. image text to text model by vllm-project · transformers

    Stale Hugging Face Hub, vllm-project 23 ♥otherUpdated 12 days ago

    vllm-project/Kimi-K3.1-Pro-pruned75
  11. project-maya

    AI repositories

    Project Maya - run GLM-5.3-Flash (321B MoE), DeepSeek-V4.1-Flash-1017 (748B MoE) and MiniMax-M3.1-Flash (428B MoE) privately on your own GPU(s), from one card to 16, NVIDIA or AMD, Linux or Windows. Engine, server, dashboard, OpenAI- and Anthropic-compatible API, quant tools.

    Stale GitHub, mw00 2,482 ★MITUpdated 13 days ago

    git clone https://github.com/mw00/project-maya.git
  12. llama-modchip

    Agents & frameworks

    Run 284B DeepSeek-V4-Flash (<59GiB RAM: 41.1 cold pp, 9.1 N=1 tg, up to 25.3 spec tok/s) & 30B MoEs (22.4 N=1 / 111.6 spec tok/s) on 64GB CPU (0 GPUs).

    Indexed GitHub, synchopate 1,137 ★Apache-2.0Updated 4 days ago

    git clone https://github.com/synchopate/llama-modchip.git
  13. janas

    AI repositories

    37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.

    Indexed GitHub, prabanta-dev 717 ★GPL-3.0Updated yesterday

    git clone https://github.com/prabanta-dev/janas.git
  14. PolyStrata

    AI repositories

    Run large AI models on your own PC: Qwen3.8-Flash-Next on Strata's engine, and GLM 5.3 Flash, DeepSeek V4 Flash, LFM2.5, Qwen3.6, Ornith 1.5, Gemma 4, Tiel Coder and MiniCPM-V on a second engine written for PolyStrata. One-click install for Windows and Linux, OpenAI/Anthropic API on localhost.

    Stale GitHub, VecSzn 722 ★MITUpdated 17 days ago

    git clone https://github.com/VecSzn/PolyStrata.git
  15. slotstream

    AI repositories

    Run a 105 GB AI model on a Mac that can't hold it. Qwen3.8-Flash-Next 4-bit, streamed from SSD on 16–64 GB Apple Silicon Macs. Fully native Swift, MLX and Metal. Offline, no Python. CLI, Swift library, and Ollama/OpenAI/Anthropic APIs. Works with Claude Code and Codex.

    Indexed GitHub, carloslfu 673 ★MITUpdated yesterday

    git clone https://github.com/carloslfu/slotstream.git
  16. imp

    Agents & frameworks

    Run large AI chat models privately on your own NVIDIA RTX 50 or RTX PRO graphics card. On one RTX 5090, Qwen3.8-27B writes about 70 words per second, 1.4x faster than vLLM. Even a model almost four times larger than the card's memory runs. Works with apps made for ChatGPT or Claude. Starts with one docker command.

    Stale GitHub, kekzl 124 ★MITUpdated 13 days ago

    git clone https://github.com/kekzl/imp.git
  17. Ling 3.1 Flash is a hybrid reasoning mixture-of-experts model from inclusionAI, with 25B active parameters out of 560B total.

    Indexed OpenRouter model list, inclusionai Updated 46 days ago

    inclusionai/ling-3.1-flash
  18. DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.

    Indexed OpenRouter model list, deepseek Updated 3 months ago

    deepseek/deepseek-v4-pro-0813
  19. Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...

    Indexed OpenRouter model list, tencent Updated 2 months ago

    tencent/hy4-preview
  20. Hy3 preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels across disabled, low, and high modes, allowing it to...

    Indexed OpenRouter model list, tencent Updated 6 months ago

    tencent/hy3-preview
  21. Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass. It is optimized for...

    Indexed OpenRouter model list, moonshotai Updated 16 months ago

    moonshotai/kimi-k2
  22. Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tasks. Its unique...

    Indexed OpenRouter model list, qwen Updated 18 months ago

    qwen/qwen3-30b-a3b
  23. Tencent: Hy3

    Models

    Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...

    Indexed OpenRouter model list, tencent Updated 4 months ago

    tencent/hy3
  24. North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized...

    Indexed OpenRouter model list, cohere Updated 5 months ago

    cohere/north-mini-code:free

Indexed entries aren't reviewed or tested by InnerVault. Details come straight from each source. Anything that can reach your accounts, files or browser can do damage, so read the source before you install it.