OSS Today

Showing 30 of 5,118 · Category: LLM

601
ziuus/goroute

⚡ High-Performance LLM Gateway in Go + Rust SIMD (< 5MB RAM) • Multi-Account Pooling • 338+ Providers • Autonomous Cascades

Go★ 70.6/day
602
Anil-matcha/awesome-uncensored-coding-models

Muapi-hosted coding-model benchmark and setup guides for Codex, Claude Code, and OpenCode; compare coding quality and refusal behavior.

Python★ 70.3/day
603
DarkHater01/tokentax

Measures how much of an AI agent's token bill is waste. tax = tokens spent / tokens needed, ranked per model x harness. Zero deps.

JavaScript★ 70.5/day
604
JustVugg/flint

Help AI agents use fewer tokens and keep more useful context. Reversible code, JSON and text compression, persistent memory, focused reads and version-aware edits.

Python★ 71/day
605
aranculleton/last-job-search

An AI agent system on Claude Code that runs your whole job search: finds and scores roles, tailors and fact-checks your resume, finds who to message on LinkedIn, and reminds you to follow up.

Python★ 70.5/day
606
hemanth1999k/mistrialNEW

Is your LLM judge flaky? Re-run any evaluator across runs, judge models and rubric wordings, and get a reliability card.

Python★ 72.7/day
607
jangviktor-web/ye-tianshi-skill

叶天士 · 温病卫气营血辨证 skill(AI agent 角色技能包 / 中医思维蒸馏 / prompt persona):把清代医家叶天士的辨证坐标系、六套框架、七条用药之禁与文人医家口吻,蒸馏为可调用能力。含《温热论》《临证指南医案》《叶氏医案存真》《叶选医衡》四部原著 10338 行正典、17 份维度分析、215 案医案库、158 条古今病名映射、20 组剂量卡、29 问 FAQ、校勘日志与可重跑核验脚本。引文逐字可核,不拟方、不诊断、不替代就医。温病派 · 病证条辨 · 读案教学 · 医学史 | Ye Tianshi TCM agent skill corpus · 卫气营血 · 中医知识库 · 经方派请用倪海厦 · 温病四大家之首 · 公版古籍数字化 · 不替代就医 · 医学教育

Python★ 70.3/day
608
muzaffermahi/rumi

A quiet place to write, with agents working beside your words. Local-first, built on Pi.

JavaScript★ 72/day
609
adityajha2005/decisionkit

Fast, cheap, probability-scored decisions for coding agents, scripts and CI. CLI + MCP server + Claude Code plugin for the OpenAI Decisions API and Jev (TypeSafe AI).

TypeScript★ 70.4/day
610
iammrduncan/thermite

Auto researching harness that burns through everything.

TypeScript★ 70.4/day
611
sedthh/batch_and_vibe

Black & White LLM Creature Mod connects an LLM to your creature’s thoughts and actions, letting you chat and command it, while also preserving the game’s ORIGINAL AI

Python★ 70.5/day
612
nettanvirdev/inertia

inertia-ai: a desktop workspace for AI agents. Chat, built-in tools, MCP and OpenAPI integrations, Docker/Daytona sandboxes with a live desktop, routines and memory. Bring your own model and keys.

Rust★ 70.3/day
613
shike/fde.dropleap.cn

中国最权威的 FDE(前置部署工程师)名录 — fde.dropleap.cn

JavaScript★ 70.3/day
614
IbrahimSobh/AISkills

A collection of useful AI skills (SKILL.md) for Claude and other AI agents, starting with a sourced fact-checker.

HTML★ 70.3/day
615
OneHillAI/Anthill

Your own AI, owned, not rented. Runs open models on your machine or your cloud, builds a wiki it maintains itself, and trains a model you keep.

Python★ 70.2/day
616
NeoPanic/harness3

Harness³ — A composable Rust harness platform for agent collectives. Durable sessions, pluggable tools, policies, and context management.

Rust★ 70.9/day
617
Romandredan/odata1c-gate

Local MCP gateway to 1C:Enterprise OData with a pseudonymisation gate: AI agents query a live 1C base, but tax IDs, accounts, phones and names reach the model only as tokens. · MCP-шлюз к OData 1С с гейтом псевдонимизации: агент работает с живой базой, а ИНН, счета, телефоны и названия видит только токенами.

Python★ 70.2/day
618
ratishjain12/ai-engineering-notes

Learning AI engineering in public, one topic at a time.

JavaScript★ 70.2/day
619
umouyoumou-del/DeepInfer-SM86NEW

DeepInfer SM86 native Qwen2.5 kernels plus a local OpenAI-compatible productivity gateway

Python★ 72.6/day
620
sametcelikbicak/opencode-team

AI agent team that takes an idea from research to launch — 14 roles, 12 pipeline stages, built on opencode. No framework, no extra dependencies.

PowerShell★ 71.1/day
621
iMaatin/Mentat

A two-layer coding agent. Your code and data never leave your machine.

TypeScript★ 70.4/day
622
parth-thakre/exl3-rdna4-recipe

Qwen3.8-27B (EXL3) on an AMD RX 9070 XT: ~150 tok/s for code, 96k context, one-line install (container, any Linux distro)

Shell★ 70.3/day
623
183600/xuen

随机灵感词 × 外部 agent 模拟研究者的 TUI 循环

Python★ 70.3/day
624
pratikgx/local-ai-field-notes

Notes from running AI on my own Mac mini: model speeds, what broke, the benchmark script, and people building local AI.

Python★ 70.4/day
625
Alekla0126/alchemist-coder

Free, open-source desktop app for Claude Code and Codex: follow every agent and subagent live, run your own agent organization and automate routines. Also searches your Gemini CLI and Grok history. macOS, Windows, Linux.

TypeScript★ 70.2/day
626
marvinli001/wevir

One local-first desktop workspace for every AI coding agent — Codex, Claude, OpenCode and your own models.

TypeScript★ 70.5/day
627
MatiasV3B/Autono

⚡ Autonomous AI Browser Agent & Navigation Copilot (Manifest V3) with Google Gemini, Anthropic Claude & OpenAI

JavaScript★ 70.3/day
628
arielshakaramiro/trending-ai-repos

Daily tracker of AI repositories trending on GitHub, with weekly analysis of rising topics and dominant languages, charts and CSV data, auto-committed by GitHub Actions.

Python★ 70.3/day
629
kldhsh123/Afterglow-Desktop

基于 RAG(免微调)重现语气与记忆的 Afterglow 桌面端。让曾经对话的温度,延续为往后的陪伴。 | Official Desktop client for Afterglow, powered by RAG (no fine-tuning needed). Let the warmth of past conversations continue into companionship.

Python★ 71.6/day
630
xisnn/dsh-outpost

DeepSeek Harness 的非官方客户端:Android App(凭据存 Keystore、发送重试不重复执行)+ Windows 桌面助手 / Unofficial DeepSeek Harness client: Android app + Windows desktop tray helper.

JavaScript★ 70.7/day