Vector Embedding Storage Calculator (Vector Database Size)
Enter how many vectors and their dimensions to size a vector database.
Your numbers
Your results
Vector Embedding Storage Calculator (Vector Database Size) by TheCalcTools
How the Vector Embedding Storage Calculator (Vector Database Size) works
Each embedding takes dimensions × bytes per number. A million 1,024-dimensional float32 vectors take about 4.1 GB before indexing.
Approximate nearest-neighbour indexes such as HNSW add graph links on top, often 30 to 100% more. Store the source text and metadata too, or keep them elsewhere and store only IDs.
The formula
Worked example
Using the default values shown in the calculator:
| Number of vectors (chunks) | 1e+06 |
|---|---|
| Embedding dimensions | 1024 |
| Precision | float32 (4 bytes) |
| Metadata and text per vector (bytes) | 1000 |
| Index overhead (e.g. HNSW graph) | 50 % |
- 1,000,000 × 1,024 × 4 bytes = 4.10 GB of vectors.
- With 50% index overhead and 1 KB of text each: 7.14 GB.
Result:
| Total storage (GB) | 7.14 |
|---|---|
| Raw vectors (GB) | 4.10 |
| Metadata and text (GB) | 1.00 |
| Bytes per vector | 7,144 |
Frequently asked questions
How can I reduce vector storage?
Use fewer dimensions (some models support shortened embeddings), lower precision (float16, int8 or binary quantisation), or product quantisation in the index.
Sources and further reading
Related calculators
More in AI & Computing · All calculators
This calculator gives estimates for information only. Check important results independently before you rely on them. See our disclaimer.