store

STORE: upsert a vector for semantic retrieval, scoped to...

STORE: upsert a vector for semantic retrieval, scoped to YOUR wallet. Bring your own embedding (any model, any dimension up to 4096) — store it under an id with optional body text + JSON meta, then find nearest neighbors with store.vector-query. Re-upserting the same id replaces it. This is the memory/RAG primitive for agents: no embedding key required on our side, no account, pay per call with x402. Kept on a rolling 90-day window.

post/api/store/vector-upsert

Request

  • Base URL: https://2s.io
  • URL: https://2s.io/api/store/vector-upsert
  • Auth: API key in header PAYMENT-SIGNATURE

Request body

nsstring required

Namespace (an index) to group vectors. Vectors in a namespace should share dimensionality.

idstring required

Vector id within the namespace, up to 256 chars.

embeddingnumber[] required

The embedding as a float array (your model's output). 1-4096 dims.

bodystring

Optional source text returned with query hits (e.g. the chunk this vector embeds).

metaunknown

Response

(unresolved $ref)

Changes

No recorded changes to this operation across all 1 revision of this API.