Search

Reusable LLM Routing Library - Centralized routing logic for model selection across applications

A debugger for RAG applications: diagnoses retrieval, context, and grounding failures.

発言が詭弁かどうかを4行で即答する Claude Code スキル。詭弁の特徴のガイドライン19条＋論理学の誤謬で判定し、本来の論点と論点に戻す一言まで返す

AI-powered image and video background removal using BiRefNet GGUF model — outputs transparent PNG/MOV/WebM

Server-side image compression skill with CDN output — supports batch processing and single-file mode

Wrapper around v8's Profiler APIs used to gather coverage data for CircleCI's Smarter Testing Javascript plugins.

Apex Copilot CLI — AI advisor for Web3 founders

Pi extension exposing the tickets daemon (GitHub/GitLab/Jira issue tracking) as an LLM-callable tool.

Official TypeScript SDK for Mesh API. One OpenAI-compatible key, 321 LLMs, unified billing, prompt templates, batches.

Swifty — CLI coding agent

CLI for installing and operating the Base Site Agent Skill.

AI-powered browser automation agent. Chat with ChatGPT from your terminal via Puppeteer automation.

No description available

No description available

pipeline-cli — the single subcommand-router home all pipeline tooling folds into; `pipeline-cli <tool> …` dispatches to a registered tool (epic #994)

ACP Next — model-driven context management for OpenCode, based on opencode-acp and DCP

Audit Pi system prompt, tool schema, context usage, and MCP overhead.

Query-aware model routing extension for Pi. Select models by task complexity, cost, context, speed, or configured order.

Cost-performance management for a stable of AI models: meter every call by source-of-funds, benchmark the market by use case, and let a procurement agent test and adopt cheaper-or-better models — reversibly and fully logged.

Official Stevai SDK — call the Stevai network from JavaScript and TypeScript. No third-party AI SDK required.