# vector-graph-rag

> Graph RAG with pure vector search, achieving SOTA performance in multi-hop reasoning scenarios.

- **URL**: https://www.freshcrate.ai/projects/vector-graph-rag
- **Author**: zilliztech
- **Category**: Databases
- **Latest version**: `v0.1.3` (2026-04-08)
- **License**: MIT
- **Source**: https://github.com/zilliztech/vector-graph-rag
- **Homepage**: https://zilliztech.github.io/vector-graph-rag/
- **Language**: Python
- **GitHub**: 66 stars, 11 forks
- **Registry**: github
- **Tags**: `graph-rag`, `graphrag`, `knowledge-base`, `knowledge-graph`, `milvus`, `multihop-question-answering`, `python`, `rag`, `sota`, `vector-database`

## Description

Graph RAG with pure vector search, achieving SOTA performance in multi-hop reasoning scenarios.

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v0.1.3` | 2026-04-08 | High | ## What's Changed  ### Bug Fixes - Replace `eval()` with `json.loads()` in NER cache loading to eliminate code injection risk (#23) - Apply retry decorator to OpenAI embedding API calls (was created but never used) (#23) - Fix hardcoded embedding fallback dimension (1536 → dynamic detection) (#23) - Use `__version__` for API version instead of hardcoded `"0.1.0"` (#23)  ### Documentation - Restructure Milvus backend section in README - Add frontend screenshot to README  **Full Changelog**: https |
| `v0.1.2` | 2026-03-26 | Medium | Tag v0.1.2 |
| `v0.1.1` | 2026-02-10 | Low | Tag v0.1.1 |
| `v0.1.0` | 2026-02-09 | Low | Tag v0.1.0 |

## Citation

- HTML: https://www.freshcrate.ai/projects/vector-graph-rag
- Markdown: https://www.freshcrate.ai/projects/vector-graph-rag.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/vector-graph-rag/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._