Ingest local files into Qdrant vector database
Ingest local documents into Qdrant for semantic search. Use this to index PDFs, Markdown, code, and other files so you can search them with `qdrant-find`.
/plugin marketplace add mindmorass/reflex/plugin install reflex@mindmorass-reflex<path> [--collection NAME] [--chunk-size WORDS] [--recursive]Ingest local documents into Qdrant for semantic search.
uvx --with pymupdf,fastembed,qdrant-client,python-docx,ebooklib,beautifulsoup4 \
python ${CLAUDE_PLUGIN_ROOT}/scripts/ingest.py <path> [options]
| Option | Default | Description |
|---|---|---|
--collection NAME | personal_memories | Qdrant collection name |
--chunk-size N | 400 | Target words per chunk |
--recursive | false | Recursively process directories |
--qdrant-url URL | http://localhost:6333 | Qdrant server URL |
| Format | Extensions |
|---|---|
.pdf | |
| Markdown | .md, .markdown |
| Text | .txt, .rst |
| HTML | .html, .htm |
| EPUB | .epub |
| Word | .docx |
| Jupyter | .ipynb |
| Mermaid | .mmd, .mermaid |
| Code | .py, .js, .ts, .go, .rs, .java, .c, .cpp, .rb, .sh, .sql, .yaml |
# Single PDF
/reflex:ingest ~/Documents/manual.pdf
# Markdown notes
/reflex:ingest ~/notes/decisions.md
# Directory of docs
/reflex:ingest ~/Documentation/ --recursive
# Custom collection
/reflex:ingest ~/research/paper.pdf --collection research_papers
# Larger chunks for context
/reflex:ingest ~/book.epub --chunk-size 600
# Mermaid architecture diagrams
/reflex:ingest ~/diagrams/ --recursive --collection architecture_patterns
Search ingested content:
qdrant-find: "how to configure authentication"
Filter by source:
qdrant-find with filter:
source: "local_file"
filename: "manual.pdf"