RAG · Search

This repo open-sources a full Chinese textbook on AI agent engineering—complete with Markdown source, compiled PDF, and runnable Python demos for every chapter.

wigolo gives AI coding agents a local, keyless web research layer that runs entirely on your machine and bills nothing per query.

Moss exists because calling out to a remote vector database adds 200–500 ms of latency—enough to kill a real-time conversation—so it runs embedding and search inside your process instead.

Heym exists because serious AI automation shouldn't require a sales call to access observability, evals, or multi-agent orchestration.

A self-hosted framework for building real-time voice-first AI agents that persist memory, delegate long tasks to background sub-agents, and optionally show up as lip-synced digital humans.

Vector search shouldn't force you to choose between PostgreSQL and DuckDB.

Macro exists because context dies when your email, chat, docs, and tasks live in separate apps, so it links them into one queryable workspace with shared AI memory.

turbovec exists so you can index embeddings immediately—no training, no tuning, no rebuilds—and search them faster than FAISS in a fraction of the RAM.

Turns URLs into clean markdown and JSON so AI agents don't have to parse HTML soup.

A CLI generator that spits out complete FastAPI + Next.js AI apps with agents, RAG, auth, and billing wired in.

It exists because RAG shouldn't be the only way to turn a file dump into usable knowledge.

AutoFlow turns documentation into a conversational knowledge graph, then lets you embed the chat window anywhere.

MEMANTO treats memory as an active agent, not a passive vector dump, with typed storage and instant retrieval.

LycheeMemory gives LLM agents persistent, structured long-term memory across runtimes without the usual token bloat or graph-database baggage.

memsearch gives Claude Code, Codex CLI, and other agents a unified, persistent memory layer backed by Markdown and Milvus.

PixelRAG renders documents into screenshot tiles and retrieves them visually, preserving tables and layout that HTML parsers strip away.

A pipeline that turns messy PDFs and slides into structured, navigable memory for AI agents instead of flat text shards.

It exists because builders wanted a Claude Cowork-style agent that keeps work context local, encrypted, and auditable instead of shipping memory to someone else’s server.

An open-source context layer that stops AI agents from guessing what your business data means.

Bisheng is an open LLM devops platform that bundles agent orchestration, RAG, and high-precision document parsing for teams who measure readiness in RBAC rules, LDAP integrations, and the ability to pause an agent mid-flight.