GH·RADAR

Simulator comparing cache-aware routing strategies for LLM serving across GPU replicas

McLang - 用麦当劳菜单写代码的编程语言 | An esoteric programming language whose instructions are real McDonald's menu items. Turing-complete, zero dependencies, and it prints you a calorie receipt.

End-to-end inbound lead automation: webhook intake, Claude qualification (structured outputs), deterministic routing, HubSpot/Slack sync, checkpointed retries, dead-letter queue, human review queue, evals, and measured before/after.

Top AI Gateway Routing (Opensource) Star if you like it!

Bounded HTTP/HTTPS readiness checks with concurrency, cancellation, JSON assertions and metadata-only CI reports. No embedded credentials or npm dependencies.

A self-hosted URL shortener with QR codes and click tracking. Built with Express, PostgreSQL, and React.

A Telegram bot and Mini App for the customers of Remna Control: subscriptions, balance, referrals, tariffs and payments (Telegram Stars, YooKassa, CryptoBot and more). Built with Go and React.

OpenCode con el harness de gentle-shell portado: ODD, subagentes enfocados, review por riesgo, memoria persistente y guardas de seguridad.

为 Reeden 接入自己的 AI 阅读伙伴：OpenAI 兼容接口、客户端工具往返、流式回复与 SQLite 共享阅读记录。

One generated MCP server over four local apps: 50 curated tools, server-side impact-token guardrails, trifecta audit

Local planning tool enforcing a Plan→Build→Test→Review protocol with API-level refusal gates

OpenStackClient plugin for Ferry: the openstack backup commands for backup, restore and disaster recovery

酷安只读命令行工具：搜索动态/用户/应用、动态正文与评论区、用户主页

Local Camera Vision — a privacy-conscious, self-hosted camera pipeline: RTSP capture, YOLO detection, optional local VLM classification, and a small event dashboard. Runs in safe mock mode with no cameras.

Zero-dependency CLI for static environment variable drift detection, secret leak scanning, and .env.example sync.

Terse — an AI-native code language: a machine-designed, token-dense format for AI coding agents. Open source, Apache-2.0.

Remote MCP server for x402 agent payments and ERC-8004 identity analytics (streamable HTTP)

Estimate chunk read amplification against your actual slice workload.

Inspect a Wasm module’s host requirements without instantiating it.

Rehearse a migration on synthetic data and inspect what survives.

Self-hosted POS with atomic inventory, partial refunds and cash reconciliation

Load benchmark for OpenAI-compatible LLM servers with character-chat and RAG-shaped traffic: cost per 1M output tokens, per turn and per session at your latency SLO (built on AIPerf)

A limited-stock pickup shop with expiring reservations and a private order desk

Safe file editing for AI agents: back up, validate, roll back on failure. Zero-dependency MCP server.