ollama

node-ollama

Interface with an ollama instance over HTTP.

Table of Contents

• Install
• Usage
• API
  • Ollama
  • generate
  • create
  • tags
  • copy
  • delete
  • pull
  • embeddings
• Building
• Testing

Install

npm i ollama

Usage

import { Ollama } from "ollama";

const ollama = new Ollama();

for await (const token of ollama.generate("llama2", "What is a llama?")) {
	process.stdout.write(token);
}

API

The API aims to mirror the HTTP API for Ollama.

Ollama

new Ollama(config);

• config <Object> The configuration object for Ollama.
  • address <string> The Ollama API address. Default: "http://localhost:11434".

Create a new API handler for ollama.

generate

ollama.generate(model, prompt);

• model <string> The name of the model to use for the prompt.
• prompt <string> The prompt to give the model.
• Returns: <AsyncGenerator<string, GenerateResult>> A generator that outputs the tokens as strings.

Generate a response for a given prompt with a provided model. The final response object will include statistics and additional data from the request.

create

ollama.create(name, path);

• name <string> The name of the model.
• path <string> The path to the Modelfile.
• Returns: AsyncGenerator<CreateStatus> A generator that outputs the status of creation.

Create a model from a Modelfile.

tags

ollama.tags();

• Returns: Promise<Tag[]> A list of tags.

List models that are available locally.

copy

ollama.copy(source, destination);

• source <string> The name of the model to copy.
• destination <string> The name of copied model.
• Returns: Promise<void>

Copy a model. Creates a model with another name from an existing model.

delete

ollama.delete(model);

• model <string> The name of the model to delete.
• Returns: Promise<void>

Delete a model and its data.

pull

ollama.pull(name);

• name <string> The name of the model to download.
• Returns: AsyncGenerator<PullResult> A generator that outputs the status of the download.

Download a model from a the model registry. Cancelled pulls are resumed from where they left off, and multiple calls to will share the same download progress.

embeddings

ollama.embeddings(model, prompt);

• model <string> The name of the model to generate embeddings for.
• prompt <string> The prompt to generate embeddings with.
• Returns: Promise<number[]> The embeddings.

Generate embeddings from a model.

Building

To build the project files run:

npm run build

Testing

To lint files:

npm run lint