Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

SignalAdvertise
DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats

Busbar
Busbar
Omnigraph
Omnigraph
Bifrost
Bifrost
Websitebench
Websitebench

Niko1221/Strata

Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

#LLM inference
Visit GitHub
C++
44.4k
4.9k
1 contributors
MIT License
AchievementsView all →
FeaturedAdvertise to 180K+ monthly visitors
Busbar

Control what AI can do before it acts

Omnigraph

Versioned graph database on object storage. Git-style workflows and multimodal retrieval

Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

Websitebench

Can AI agents reverse-engineer real websites?

Momentum

How fast Niko1221/Strata is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Sep 23 – Nov 15, 2026

Stars +44,427 in 54 days

01K2K3K
2,921

Forks +4,945 in 54 days

0100200300
210

Merged PRs 30 in 54 days

051015
12

Trendshift Trendingwhere the repository ranked each day it made the top 25

Top 25 overall

151231414623

Top 25 in C++

22311111111136532435344632143222121132331233444756
Sep 23Sep 30Oct 7Oct 14Oct 21Oct 28Nov 4Nov 11
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRsAll languagesC++
Wed, Sep 23, 2026+0+00Not trendingNot trending
Thu, Sep 24, 2026+8+00Not trendingNot trending
Fri, Sep 25, 2026+21+40Not trending#22
Sat, Sep 26, 2026+140+195Not trending#3
Sun, Sep 27, 2026+303+294Not trending#1
Mon, Sep 28, 2026+586+629#15#1
Tue, Sep 29, 2026+649+8612#12#1
Wed, Sep 30, 2026+1,313+970#3#1
Thu, Oct 1, 2026+1,660+1160#1#1
Fri, Oct 2, 2026+1,601+1460#4#1
Sat, Oct 3, 2026+2,188+1850#1#1
Sun, Oct 4, 2026+2,462+2100#4#1
Mon, Oct 5, 2026+2,921+2010#6#1
Tue, Oct 6, 2026+1,837+1780#23#3
Wed, Oct 7, 2026+1,326+1630Not trending#6
Thu, Oct 8, 2026+1,386+1360Not trending#5
Fri, Oct 9, 2026+1,100+1280Not trending#3
Sat, Oct 10, 2026+1,163+1510Not trending#2
Sun, Oct 11, 2026+907+1170Not trending#4
Mon, Oct 12, 2026+1,118+1700Not trending#3
Tue, Oct 13, 2026+1,001+1350Not trending#5
Wed, Oct 14, 2026+891+1500Not trending#3
Thu, Oct 15, 2026+856+1250Not trending#4
Fri, Oct 16, 2026+799+1000Not trending#4
Sat, Oct 17, 2026+646+1060Not trending#6
Sun, Oct 18, 2026+564+770Not trending#3
Mon, Oct 19, 2026+824+940Not trending#2
Tue, Oct 20, 2026+919+900Not trending#1
Wed, Oct 21, 2026+736+1060Not trending#4
Thu, Oct 22, 2026+669+870Not trending#3
Fri, Oct 23, 2026+563+800Not trending#2
Sat, Oct 24, 2026+822+830Not trending#2
Sun, Oct 25, 2026+771+900Not trending#2
Mon, Oct 26, 2026+955+1040Not trending#1
Tue, Oct 27, 2026+846+980Not trending#2
Wed, Oct 28, 2026+760+840Not trending#1
Thu, Oct 29, 2026+664+940Not trending#1
Fri, Oct 30, 2026+518+680Not trending#3
Sat, Oct 31, 2026+580+640Not trending#2
Sun, Nov 1, 2026+609+1040Not trending#3
Mon, Nov 2, 2026+739+830Not trending#3
Tue, Nov 3, 2026+858+690Not trending#1
Wed, Nov 4, 2026+658+1040Not trending#2
Thu, Nov 5, 2026+612+760Not trending#3
Fri, Nov 6, 2026+515+630Not trending#3
Sat, Nov 7, 2026+427+520Not trending#4
Sun, Nov 8, 2026+509+530Not trending#4
Mon, Nov 9, 2026+495+490Not trending#4
Tue, Nov 10, 2026+367+490Not trending#7
Wed, Nov 11, 2026+362+520Not trending#5
Thu, Nov 12, 2026+316+520Not trending#6
Fri, Nov 13, 2026+367+390Not trendingNot trending
Sat, Nov 14, 2026+282+430Not trendingNot trending
Sun, Nov 15, 2026+238+240Not trendingNot trending
Total+44,427+4,94530
  • 1#1 on the list
  • 7Rank that day
  • Not trending
  • Mentions and achievements

    Where people are talking about Niko1221/Strata, and every badge it has earned
    1. Nov 151 post
      • Same local Qwen 3.8 IQ3_XXS, same prompt: build a Three.js voxel kingdom. Strata / Claude Windows app : 35m 32s across build + polish. Strata / Trae: 2h 11m across a failed run + recovery. Final…

        Read full on X, posted by @blue0orbit

        @blue0orbit · x.com

    2. Nov 143 posts
      • got qwen3.8 flash-next running on a single intel arc pro b60. 24gb card, windows. 35-45 tokens/sec c2, full 262k context on each abliterated with priority lanes it can give one stream priority so it…

        Read full on X, posted by @HarshKamdar67

        @HarshKamdar67 · x.com

      • New RTX 5090 32GB Strata benchmark: Qwen3.8-Flash-Next IQ3_S reached 173.6 tok/s in short chat and 222.9 tok/s in agent follow-ups. 16K prefill: 6,674 tok/s. Single-system, speed-only test; no…

        Read full on X, posted by @zzz1367761

        @zzz1367761 · x.com

      • 想在自己电脑上跑大模型,又怕内存不够? Strata 是免费开源的本地运行工具,桌面应用和终端都能用,不用账号。 内存放不下的部分会从 SSD 读取,所以对内存要求低一些,代价是要留足磁盘空间。 DeepSeek V4.1 Flash 这类模型都能跑。 地址:

        Read full on X, posted by @Antonio17495254

        @Antonio17495254 · x.com

    3. Nov 111 post2 posts
      • Best stack I've found for building projects with local AI agents: @papercliping — Your AI company. Hire agents, set goals, let them build. Strata — Blazing-fast local inference. Ponytail — No-BS…

        Read full on X, posted by @RogerBronchal

        @RogerBronchal · x.com

      • RTX 5080 + 4090: a Strata user traced a mixed-GPU prefill regression to VRAM budgeting on the wrong device. A one-line local fix raised 32K prompt throughput from 960 to 3,411 tok/s. 10/10 local runs…

        Read full on X, posted by @zzz1367761

        @zzz1367761 · x.com

    4. Nov 102 posts
      • This week’s chatter is about running large MoE models on a regular machine. Read the repos and test on your own hardware before trusting the headline numbers. Strata: Colibri:

        Read full on X, posted by @openreponotes

        @openreponotes · x.com

      • Fine-tuning a 125B quantized model on one 48GiB GPU? A new Strata PR reports a zero-order training path adding only ~0.40GiB VRAM over inference for short samples. Pipeline tested; learning quality…

        Read full on X, posted by @zzz1367761

        @zzz1367761 · x.com

    5. Nov 92 posts
      • Strata 0.2.12, 7 Nov. One solo dev. Runs Alibaba's Qwen3.8-Flash-Next, 125B active params, on a 12GB card. RTX 3060 class: 87 to 264 tokens a second at 2K context. Card you own. No seat, no renewal.…

        Read full on X, posted by @BareStack

        @BareStack · x.com

      • 最近看到个开源项目 Strata,挺有意思。 普通电脑也能跑大模型,还能接入 Claude Code、Codex。 准备拿自己的电脑折腾一下,看看实际效果到底怎么样。 项目:

        Read full on X, posted by @lovingkidzs

        @lovingkidzs · x.com

    You might also like

    Repositories close to Niko1221/Strata
    FeaturedAdvertise to 180K+ monthly visitors
    Bifrost

    Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

    Websitebench

    Can AI agents reverse-engineer real websites?

    Omnigraph

    Versioned graph database on object storage. Git-style workflows and multimodal retrieval

    Busbar

    Control what AI can do before it acts

    mach10x/strata

    Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

    21222

    architectds/Strata

    C++

    Qwen3.8-Flash-Next (125B MoE) on a 8GB+ NVIDIA GPU: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

    887

    hokar3361/Strata-Temp-MacPatchFork

    C++

    Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

    241

    MistVVK/XeStrata

    C++

    Qwen3.8-Flash-Next (125B MoE) on a 8GB+ Multi Vendor GPU 🥳No CUDA!

    35425

    simplexai-labs/PolyStrata

    C++

    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.

    1.1k143

    eddoursul/Lamina

    C++

    A fork of https://github.com/Niko1221/Strata optimized for my 3090+5070 Ti setup, see docs/COMPARISON.md. Renamed to Lamina.

    635

    robertkarl/UnbiasedStrata

    C++

    Strata minus the bullshit. clone, rsync to prod, compile and go

    23758

    DevDock-AI/Speedy

    C++

    Qwen3.8-Flash-Next (125B MoE) up to 1M context on one or two RTX cards (tuned on 3080 10gb). A measured, tuned and benchmarked fork of Strata.

    773

    jmnargi/Strata-V100

    C++

    Strata port for Tesla V100 (sm_70): Qwen3.8-Flash-Next with 262K context and MTP

    15020