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.
mongodb/voyage-4Runs on your balance at the prices above. New accounts start with free usage.
Settings
VoyageAI by MongoDB: voyage-4 is an embedding model from VoyageAI by MongoDB. It belongs to the embeddings category, which means its job is to turn text into vector representations rather than to generate chat responses. It is a general-purpose model, and it is built to handle multilingual content as well as English. If your application depends on semantic similarity, retrieval, or search quality, this is the class of model you would evaluate.
The model accepts text as input and returns embeddings as output. Its context length is 32,000 tokens, and its maximum output is 28,800. Because the output is a vector rather than prose, you will not use it for conversation, summarization, or code generation. Instead, you send it text and receive the numeric representation that downstream systems use for comparison and ranking. The multilingual support means the same model can be used across content in different languages without switching to a separate embedding model for each one.
The practical uses follow directly from that shape. Retrieval and search are the primary targets, so voyage-4 fits semantic search over a document collection, retrieval-augmented generation pipelines where you need to fetch relevant passages before passing them to a language model, and AI applications that rely on similarity between queries and stored content. Multilingual retrieval is a natural fit as well, since queries and documents do not have to be in the same language for the embeddings to be useful. Clustering, deduplication, and recommendation-style matching are also reasonable applications, because they all reduce to comparing embeddings.
You call the model through api14's OpenAI-compatible API using the model id "mongodb/voyage-4". Because the interface follows the OpenAI format, you can point an existing OpenAI client at api14 and send embedding requests without rewriting your integration. The model lists no supported parameters beyond the standard request fields, so keep your requests simple and avoid assuming tuning options that are not part of the interface. The model id is the value you place in the request to select this specific model.
One thing to keep in mind is scope. This is an embeddings model, so if you need text generation, reasoning, or tool use, a chat or completion model is the better choice. The 32,000 token context and 28,800 maximum output describe the embedding workload, not a chat budget. Use voyage-4 when the task is representation and retrieval, and reach for a generative model when the task is producing text.
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": "mongodb/voyage-4",
"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 VoyageAI by MongoDB: voyage-4.