Luciform Research
Autonomous AI Systems
Building the infrastructure for intelligent agents. RAG architectures, knowledge graphs, and AI-powered development tools.
Meet Lucie
AI assistant built with LangGraph that can search code, explain architectures, and answer questions about Luciform Research projects.
Hybrid Search in RagForge
"Hey Lucie, could you explain me how does the hybrid search works in your ragforge framework?"
Deep dive into RagForge's hybrid search: semantic-first with BM25 boost, RRF fusion, parallel execution.
search_knowledgeget_code_samplegrep_codeAgent Infrastructure
Tools and frameworks for building autonomous AI systems
RagForge
CORE FRAMEWORKUniversal RAG agent with persistent knowledge brain. Neo4j graph database, semantic search, and MCP integration.
@luciformresearch/ragforge→
CodeParsers
AST EXTRACTIONMulti-language code parser using tree-sitter WASM. Extract functions, classes, and imports from any codebase.
@luciformresearch/codeparsers→XMLParser
LLM OUTPUTSFault-tolerant XML parser for AI pipelines. Handles malformed LLM outputs with streaming SAX API.
@luciformresearch/xmlparser→Artifacts
Real outputs from RagForge and Lucie Agent - brain search reports, ingestion logs, and conversation traces.
Brain Search: Incremental Ingestion
MarkdownSemantic search for IncrementalIngestionManager with dependency graph visualization
Brain Search: Search Service
MarkdownExploring the SearchService class with RRF hybrid search implementation
Brain Search: Detailed
MarkdownComprehensive search with full dependency tree exploration
Ingestion Pipeline Log
LogReal-time logging of file parsing, entity extraction, and embedding generation
API Ingestion Log
LogFull trace of GitHub repo cloning, tree-sitter parsing, and Neo4j ingestion
Lucie Agent Conversation
LogExcerpt from Lucie's conversation log showing tool calls and responses