Search

Self-evolving modular memory LLM framework — biological memory × formal verification × industrial IoT mesh × TUI HITL

Drop-in Claude Code Stop hook that ships per-user, per-project usage traces to Langfuse.

OpenAI-compatible inference server: Llama 3.1 8B + Whisper + Kokoro TTS exposed via ngrok

Background plan/implement/review/address pipelines for Claude Code

Text compression that uses a local GPT-2 model as a probability source.

Split text into semantically coherent, LLM-categorized chunks

ENCOURAGE

SQL 智能助手 - 自然语言转 SQL 查询工具，支持主流数据库和 LLM

Semantic search across Claude Code sessions. Find past conversations by intent, not just keywords.

Storage layer for NeoBot — SQLAlchemy + Alembic

Unified storage adapter interface for multi-cloud storage backends

File storage dependency for FastAPI

A storage provider which can fetch and store media in Amazon S3.

Query-aware semantic chunk selection under LLM context-window budgets.

Toonic — Universal TOON Format Platform for LLM-optimized file representations

Extended fork of random-address with broader coverage and batch sampling

Hierarchical KNN intent engine for OVOS, powered by IBM Granite embeddings

Compute shannon entropy from LLM responses to detect hallucinations

Compute semantic entropy from LLM responses to detect hallucinations

Auto-generated daily briefing podcast: Claude + web search → script → Polly TTS → S3 + RSS