Operations
Embeddings and models
Create vector embeddings and discover models supported by a provider endpoint.
Embeddings
import { embedding } from "any-llm-ts";
const result = await embedding({
provider: "openai",
model: "text-embedding-3-small",
input: ["First document", "Second document"],
});
for (const item of result.data) {
console.log(item.index, item.embedding);
}The result normalizes token usage and vector data:
result.provider;
result.model;
result.usage.promptTokens;
result.usage.totalTokens;You can request dimensions and an encoding format when the provider supports them:
await llm.embedding({
model: "text-embedding-3-small",
input: "Hello",
dimensions: 512,
encodingFormat: "float",
});List models
import { listModels } from "any-llm-ts";
const models = await listModels({ provider: "openai" });
for (const model of models) {
console.log(model.id, model.ownedBy);
}Or use a reusable client:
const models = await llm.listModels();Some providers expose compatible inference endpoints but no model-list endpoint. Check
metadata.capabilities.listModels before relying on discovery.
No reranking API yet
The Python source project includes reranking integrations. This TypeScript port does not expose a rerank operation yet, so it is deliberately absent from the active API documentation.