Milvus
Context / RAG · Billion-scale vector indexing for production RAG; self-hosted or Zilliz Cloud managed
At a glance
| Setup effort | Medium |
| Released | 2019 |
| Open source | Yes |
| Hosting | Cloud |
| Privacy | Configurable |
| Update mode | Real-time |
| Staleness | auto |
| Index type | HNSW |
| Index limit | Large |
| Capabilities | HNSW/IVF/DiskANN indexing, Hybrid search, Scalar filtering, GPU acceleration, Streaming inserts |
What Milvus does
HNSW/IVF/DiskANN indexing, Hybrid search, Scalar filtering, GPU acceleration, Streaming inserts
Best for
Billion-scale vector indexing for production RAG; self-hosted or Zilliz Cloud managed
Works well with
LLM Provider / Model
Integration
Agent / Orchestration
Conflicts & caveats
- Privacy conflict: Self-hosted Llama 3 (Ollama/Groq) sends code to cloud Milvus. Use local context (Continue indexing, ChromaDB, LanceDB, pgvector, Vespa self-hosted) for true privacy.
Build a full stack around Milvus — Flowpicker shows compatibility warnings before you commit.
Open the stack planner →