HoloIndex

Semantic code retrieval and project memory substrate for the FoundUps ecosystem.

Installation

pip install holoindex

Quick Start

from holoindex import HoloIndex

# Initialize with vector storage path
holo = HoloIndex(vector_path="./vectors")

# Index your codebase
holo.index_documents(roots=["./src"])

# Search
results = holo.search("authentication middleware", limit=5)
for r in results:
    print(f"{r['path']}: {r['similarity']:.2f}")

Features

• Semantic Search: Find code by meaning, not just keywords
• ChromaDB Backend: Persistent vector storage
• Sentence Transformers: State-of-the-art embeddings
• Hybrid Scoring: Combines semantic + keyword relevance

Configuration

# Use custom embedding model
holo = HoloIndex(
    vector_path="./vectors",
    model_name="all-MiniLM-L6-v2",
)

# Configure similarity threshold
results = holo.search("query", min_similarity=0.5)

License

MIT - See LICENSE for details.

Links

• FoundUps
• Documentation