We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.
amazon/titan-embed-text-v1Runs on your balance at the prices above. New accounts start with free usage.
Settings
Amazon Titan Embeddings G1 - Text is a text embedding model from Amazon, part of the Titan family of foundation models. It is a first-generation Titan text embedding model, and its main category is embeddings. Rather than generating prose, it converts text into numerical vectors that represent meaning, which makes it a building block for search, retrieval and similarity systems.
The model accepts text input only and returns embeddings as output. Each input can be up to 8,192 tokens long, which allows fairly large documents or long passages to be embedded in a single request. It returns fixed 1,536-dimension vectors, so the shape of the output is predictable and consistent across calls, which simplifies storage in a vector database and comparison between vectors. The maximum output is 7,372.
Because the model produces embeddings rather than text, the use cases follow from similarity and retrieval. You can embed a corpus of documents and embed user queries at request time, then rank documents by vector similarity to build semantic search over internal knowledge bases, documentation or product catalogs. The same approach supports retrieval augmented generation, where relevant passages are fetched and passed to a separate text generation model. Embeddings also fit clustering and topic grouping of large text collections, deduplication of near-identical content, classification by comparing a document vector to labeled examples, and recommendation or matching features that rely on semantic closeness rather than keyword overlap. The 8,192 token input limit means long articles and reports can be embedded without splitting them into many small chunks, though chunking may still be preferable when you need passage-level retrieval.
You call the model through api14's OpenAI-compatible API using the model id "amazon/titan-embed-text-v1". Because the interface follows the OpenAI embeddings convention, existing OpenAI client libraries and SDKs can point at api14's base URL and send embedding requests with that model id, which keeps integration work small if you already have an embeddings pipeline. The model lists no supported request parameters beyond the standard embedding request fields, so plan on sending input text and the model id, and handle the returned 1,536-dimension vectors directly.
Consider whether an embedding model is the right fit at all: this model does not generate text, so tasks like summarization, question answering or chat need a separate text generation model, with Titan Embeddings used alongside it for retrieval. If you need a different vector size or a newer generation of Titan embeddings, check what else api14 exposes under the Amazon provider before committing to this model id.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
const result = await client.embeddings.create({
"model": "amazon/titan-embed-text-v1",
"input": [
"The quick brown fox",
"Lorem ipsum"
]
});
console.log(result.data.map((d) => d.embedding.length));
This community has no topics yet. Open the first one about Amazon: Titan Embeddings G1 - Text.