Niko1221/Strata

More in ai-infra

Ranked by stars gained in the last 24 hours.

• 

  deepseek-ai/DeepInfer

  DeepInfer: High-performance, large-scale inference framework for DeepSeek models

  Python23k+634/dai-infra
• 

  Edge0-AI/Edge0

  Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported

  Python7.4k+510/d7% growthai-infraai-models
• 

  experientiallabs/experiential

  Experiential is the open source, zero markup gateway for BYOK, self-hosted and 1000+ marketplace models. It learns from your traffic to cut costs, recommend better models, and train a specialized model you own.

  Python13.3k+490/d4% growthai-infra
• 

  cheaper-inference/OmniRoute

  Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors

  TypeScript81.5k+361/dai-infra
• 

  Comfy-Org/ComfyUI

  The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

  Python143.3k+343/dai-infraai-models
• 

  StarTrail-org/LEANN

  [MLsys2026 Best Paper]: https://arxiv.org/abs/2506.08276. RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

  Python13.3k+213/d2% growthai-infraai-agents
• 

  mw00/project-maya

  Project Maya - run GLM-5.3-Flash (321B MoE) and DeepSeek-V4.1-Flash-1017 (748B 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.

  C++2k+201/d11% growthai-infraai-models
• 

  JustVugg/colibri

  Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

  C44.1k+170/dai-infraai-models

Star history

Weekly, from GitHub's stargazers history, over a 30-day window.

1. 30 Sep 2026Entered our index
2. 15 Oct 2026Hacker News threadread the thread· 15 days later

Stars: 1.9k when it entered our index, 30.7k on the last capture we hold (23 Oct 2026), so +28.8k in that span. That span is our own measurement window, not a claim about what happened before it.

It has never shown up on GitHub Trending while we were watching, so there is no arrival to compare against.

We hold 11 mentions for it, the oldest dated 28 Sep 2026 and the first written into our database on 16 Oct 2026. Thread dates come from the source, not from when we fetched them.

11 items so far: 11 on Hacker News, 0 on Reddit; 0 in the last 7 days, 11 in the last 30. Checked by the pipeline, which re-scans trending repositories every few hours. Last check 1 week ago.

• LFM2.5-24B-A2B: Built for fast on-device agentsHNcomment · by eyegor · 1 week ago

  Oh you should be running flash next with strata. Just go to stratas github page [0] and it will explain everything. It runs quite fast on 16+64, you should be able to get 50 tok/s or so. It's a 180b model and it feels a lot smarter than most of the other open weight options. As far as knowing "what will fit" in general that's hard to know these days. The moe models are hard to predict and run wil…

• Janas: Run LLMs larger than RAM by streaming experts from SSDHNcomment · by MadameMinty · 2 weeks ago

  Then check out those: https://github.com/prabanta-dev/janas <- this, focuses on Qwen3-Next-80B-A3B. https://github.com/Niko1221/Strata <- Qwen3.8-Flash-Next. I use this (tuned) and I cannot understate how satisfied I am. https://github.com/antirez/ds4 <- DeepSeek 4 family + more.

• Janas: Run LLMs larger than RAM by streaming experts from SSDHNcomment · by suprjami · 2 weeks ago

  Qwen 3? (Jumanji) What year is it? What a waste of time. Strata already does this with Qwen3.8-Flash-Next: https://github.com/Niko1221/Strata

• Step 5 Preview, a 1M-context MoE from StepFun, shows up on OpenRouterHNcomment · by throwa356262 · 2 weeks ago

  Not sure why this is downvoted. It is fully possible to run the 180B Owen model on strix halo using strata and even get reasonable speed out of it. https://github.com/Niko1221/Strata

• Why isn't the industry freaking out about DeepSeek 4.1 Flash?HNcomment · by nicman23 · 2 weeks ago

  what they should freak out about is qwen 3.8 github.com/Niko1221/Strata it rips with just 64 ram and a 9070xt

Competes with

Repositories in the same collection, of a comparable size (within 5× in stars) and from a different owner, that were discussed in the same threads. Same-owner projects are excluded because they are usually pieces of the same tool, not rivals. A repository also stays out if the two never met in a discussion.

• antirez/ds424k

  Mentioned together in 3 threads — named in the same comment 1 time.

  Compare side by side
• ggml-org/llama.cpp132.1k

  Mentioned together in 1 thread.

  Compare side by side
• FlashML-org/FreeToken14.7k

  Mentioned together in 1 thread.

  Compare side by side

Reddit (fetched when this page is opened)

A Reddit search for github.com/Niko1221/Strata runs while this page is rendered, at most once every 6 hours per repository. What it finds is stored in our database (25 posts per fetch, at most 200 repositories a day), so the numbers on this site can be reproduced later. Reddit's feed has no scores or comment counts, so none are shown.

Reddit answered 429 (limit resets in 13s); next request in 11s.