Search

LLM quality plugin for InferenceBench Suite (deterministic fixture scoring; LLM-as-judge deferred)

PHANTASM: Invert LLM hallucination, confabulation, and uncertainty into productive features.

A local-first archive, search, and intelligence engine for your LLM conversation history.

LangChain tool for generating My Magic Pencil live visual lessons

CLI tool that evaluates LLM outputs from production logs against a dual-dimension rubric.

CLI tool that evaluates LLM outputs from production logs against a dual-dimension rubric.

CLI tool that evaluates LLM outputs from production logs against a dual-dimension rubric.

Python client for Membrain AI Safety Gateway — drop-in replacement for OpenAI SDK

Unified blob storage abstraction for Lexigram Framework - S3, GCS, Azure Blob, and local filesystem

Unified Efficient Fine-Tuning of 100+ LLMs

Consistent Embeddings of high-dimensional Recordings using Auxiliary variables

Detect behavioral regressions between LLM model versions

LLM-powered synthetic text data generation for text classification tasks, with multi-strategy generation, multilingual support, and quality filtering.

Local-only Python SDK mirroring PageIndex Cloud API — vectorless RAG, no cloud required.

Local usage dashboard for Codex, Claude Code, Cursor, and WakaTime AI coding time.

Local voice assistant pipeline — STT → LLM → TTS, 100% offline

Instant token count and cost estimate across all major LLMs

Convert long AI conversations into portable conversation state graphs for LLM handoffs.

A sleek Regex visualizer with enhanced Chinese support and elegant diagrams

在动手设计/开发前先判断「这是不是真需求」——基于梁宁《真需求》价值-共识-模式的 Claude Code Skill。A Claude Code skill that tells a real demand from a self-deceiving fake one before you build.