Install memista and run the server

memista ships as a crate with a bundled server binary. This guide gets you from zero to a running index with data in it.

1. Add the crate

cargo add memista

memista targets Rust 1.56+. It pulls in USearch (with the simsimd and fp16lib features) for the index and async-sqlite for metadata storage.

2. Start the server

Run the bundled binary. By default it binds to 127.0.0.1:8083 and exposes three endpoints:

• POST /v1/insert — add chunks
• POST /v1/search — nearest-neighbour query
• DELETE /v1/drop — drop a partition

memista binds to localhost and trusts everything on the wire. There is no authentication. If you expose it beyond the box, put a reverse proxy in front and terminate auth there.

3. Explore the API in the browser

The OpenAPI spec is generated by apistos and rendered four ways:

/swagger · /redoc · /rapidoc · /scalar

4. Insert and search

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":"{}"}]}'

Then query:

curl -X POST http://localhost:8083/v1/search \
  -H "Content-Type: application/json" \
  -d '{"database_id":"my_app","query":[0.1,0.2],"limit":5}'

You’ll get back ranked chunk ids with their text, metadata, and distance.

Next: insert vectors over HTTP, run a nearest-neighbour search, or read how it works.