Text Embeddings
Vector embeddings for RAG and semantic search, up to 100 texts per call.
Text embeddings for RAG / semantic search: one string or up to 100 strings -> L2-normalised vectors with open BAAI models on Cloudflare GPUs: bge-small-en-v1.5 (384-d, English, default) or bge-m3 (1024-d, multilingual). Up to 24,000 characters per call.
What agents use it for
- Index documents for retrieval-augmented generation
- Cluster or deduplicate texts by meaning
- Multilingual semantic search with bge-m3
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
input | array of string | yes | Text or list of texts (a single string is also accepted) |
model | "bge-small-en-v1.5" | "bge-m3" · default "bge-small-en-v1.5" | no | Embedding model to use |
Call it
{
"name": "embed_text",
"arguments": {
"input": [
"torque sensor calibration",
"laser wavelength stability"
],
"model": "bge-small-en-v1.5"
}
}curl -i -X POST "https://api.ishob.com/v1/embed" \
-H "Content-Type: application/json" \
-d '{"input":["torque sensor calibration","laser wavelength stability"],"model":"bge-small-en-v1.5"}'Paid with x402: the first call returns HTTP 402 with the price; an x402 client (for example @x402/fetch or @x402/mcp) signs a USDC payment and retries. Calls that fail are not charged.
Connect the MCP server: setup for Claude, Cursor and other clients.