Insert vectors over HTTP

POST /v1/insert is how data gets into memista. This guide covers the chunk shape and what happens under the hood.

The request

curl -X POST http://localhost:8083/v1/insert \
  -H "Content-Type: application/json" \
  -d '{
    "database_id": "my_app",
    "chunks": [{
      "embedding": [0.1, 0.2],
      "text": "Hello world",
      "metadata": "{\"source\": \"readme\"}"
    }]
  }'

Each chunk carries three things:

• embedding — the vector. In the stock crate this is length 2; see changing embedding dimensions.
• text — the content the vector represents.
• metadata — a string (commonly JSON) stored verbatim and returned on search.

What happens on insert

1. The row (text + metadata) is written to the SQLite table chunks_<database_id> first.
2. The returned chunk_id becomes the USearch key for the vector.
3. After the batch, the <database_id>.usearch index file is saved to disk.

Because the SQLite id and the USearch key are the same integer, search can return ranked keys and hydrate the text/metadata straight back from SQLite. See how it works for the full picture, then run a search.