// Programmatically interact with the Hub

await createRepo({
  repo: { type: "model", name: "my-user/nlp-model" },
  accessToken: HF_TOKEN
});

await uploadFile({
  repo: "my-user/nlp-model",
  accessToken: HF_TOKEN,
  // Can work with native File in browsers
  file: {
    path: "pytorch_model.bin",
    content: new Blob(...)
  }
});

// Use all supported Inference Providers!

await inference.chatCompletion({
  model: "meta-llama/Llama-3.1-8B-Instruct",
  provider: "sambanova", // or together, fal-ai, replicate, cohere …
  messages: [
    {
      role: "user",
      content: "Hello, nice to meet you!",
    },
  ],
  max_tokens: 512,
  temperature: 0.5,
});

await inference.textToImage({
  model: "black-forest-labs/FLUX.1-dev",
  provider: "replicate",
  inputs: "a picture of a green bird",
});

// and much more…

Hugging Face JS libraries

This is a collection of JS libraries to interact with the Hugging Face API, with TS types included.

• @huggingface/inference: Use all supported (serverless) Inference Providers or switch to Inference Endpoints (dedicated) to make calls to 100,000+ Machine Learning models
• @huggingface/hub: Interact with huggingface.co to create or delete repos and commit / download files
• @huggingface/mcp-client: A Model Context Protocol (MCP) client, and a tiny Agent library, built on top of InferenceClient.
• @huggingface/gguf: A GGUF parser that works on remotely hosted files.
• @huggingface/dduf: Similar package for DDUF (DDUF Diffusers Unified Format)
• @huggingface/tasks: The definition files and source-of-truth for the Hub’s main primitives like pipeline tasks, model libraries, etc.
• @huggingface/jinja: A minimalistic JS implementation of the Jinja templating engine, to be used for ML chat templates.
• @huggingface/space-header: Use the Space mini_header outside Hugging Face
• @huggingface/ollama-utils: Various utilities for maintaining Ollama compatibility with models on the Hugging Face Hub.
• @huggingface/tiny-agents: A tiny, model-agnostic library for building AI agents that can use tools.

We use modern features to avoid polyfills and dependencies, so the libraries will only work on modern browsers / Node.js >= 18 / Bun / Deno.

The libraries are still very young, please help us by opening issues!

Installation

From NPM

To install via NPM, you can download the libraries as needed:

npm install @huggingface/inference
npm install @huggingface/hub
npm install @huggingface/mcp-client

Then import the libraries in your code:

import { InferenceClient } from "@huggingface/inference";
import { createRepo, commit, deleteRepo, listFiles } from "@huggingface/hub";
import { McpClient } from "@huggingface/mcp-client";
import type { RepoId } from "@huggingface/hub";

From CDN or Static hosting

You can run our packages with vanilla JS, without any bundler, by using a CDN or static hosting. Using ES modules, i.e. <script type="module">, you can import the libraries in your code:

<script type="module">
    import { InferenceClient } from 'https://cdn.jsdelivr.net/npm/@huggingface/inference@4.13.3/+esm';
    import { createRepo, commit, deleteRepo, listFiles } from "https://cdn.jsdelivr.net/npm/@huggingface/hub@2.7.1/+esm";
</script>

Deno

// esm.sh
import { InferenceClient } from "https://esm.sh/@huggingface/inference"

import { createRepo, commit, deleteRepo, listFiles } from "https://esm.sh/@huggingface/hub"
// or npm:
import { InferenceClient } from "npm:@huggingface/inference"

import { createRepo, commit, deleteRepo, listFiles } from "npm:@huggingface/hub"

Usage examples

Get your HF access token in your account settings.

@huggingface/inference examples

import { InferenceClient } from "@huggingface/inference";

const HF_TOKEN = "hf_...";

const client = new InferenceClient(HF_TOKEN);

// Chat completion API
const out = await client.chatCompletion({
  model: "meta-llama/Llama-3.1-8B-Instruct",
  messages: [{ role: "user", content: "Hello, nice to meet you!" }],
  max_tokens: 512
});
console.log(out.choices[0].message);

// Streaming chat completion API
for await (const chunk of client.chatCompletionStream({
  model: "meta-llama/Llama-3.1-8B-Instruct",
  messages: [{ role: "user", content: "Hello, nice to meet you!" }],
  max_tokens: 512
})) {
  console.log(chunk.choices[0].delta.content);
}

/// Using a third-party provider:
await client.chatCompletion({
  model: "meta-llama/Llama-3.1-8B-Instruct",
  messages: [{ role: "user", content: "Hello, nice to meet you!" }],
  max_tokens: 512,
  provider: "sambanova", // or together, fal-ai, replicate, cohere …
})

await client.textToImage({
  model: "black-forest-labs/FLUX.1-dev",
  inputs: "a picture of a green bird",
  provider: "fal-ai",
})



// You can also omit "model" to use the recommended model for the task
await client.translation({
  inputs: "My name is Wolfgang and I live in Amsterdam",
  parameters: {
    src_lang: "en",
    tgt_lang: "fr",
  },
});

// pass multimodal files or URLs as inputs
await client.imageToText({
  model: 'nlpconnect/vit-gpt2-image-captioning',
  data: await (await fetch('https://picsum.photos/300/300')).blob(),
})

// Using your own dedicated inference endpoint: https://hf.co/docs/inference-endpoints/
const gpt2Client = client.endpoint('https://xyz.eu-west-1.aws.endpoints.huggingface.cloud/gpt2');
const { generated_text } = await gpt2Client.textGeneration({ inputs: 'The answer to the universe is' });

// Chat Completion
const llamaEndpoint = client.endpoint(
  "https://router.huggingface.co/hf-inference/models/meta-llama/Llama-3.1-8B-Instruct"
);
const out = await llamaEndpoint.chatCompletion({
  model: "meta-llama/Llama-3.1-8B-Instruct",
  messages: [{ role: "user", content: "Hello, nice to meet you!" }],
  max_tokens: 512,
});
console.log(out.choices[0].message);

@huggingface/hub examples

import { createRepo, uploadFile, deleteFiles } from "@huggingface/hub";

const HF_TOKEN = "hf_...";

await createRepo({
  repo: "my-user/nlp-model", // or { type: "model", name: "my-user/nlp-test" },
  accessToken: HF_TOKEN
});

await uploadFile({
  repo: "my-user/nlp-model",
  accessToken: HF_TOKEN,
  // Can work with native File in browsers
  file: {
    path: "pytorch_model.bin",
    content: new Blob(...)
  }
});

await deleteFiles({
  repo: { type: "space", name: "my-user/my-space" }, // or "spaces/my-user/my-space"
  accessToken: HF_TOKEN,
  paths: ["README.md", ".gitattributes"]
});

@huggingface/mcp-client example

import { Agent } from '@huggingface/mcp-client';

const HF_TOKEN = "hf_...";

const agent = new Agent({
  provider: "auto",
  model: "Qwen/Qwen2.5-72B-Instruct",
  apiKey: HF_TOKEN,
  servers: [
    {
      // Playwright MCP
      command: "npx",
      args: ["@playwright/mcp@latest"],
    },
  ],
});

await agent.loadTools();

for await (const chunk of agent.run("What are the top 5 trending models on Hugging Face?")) {
    if ("choices" in chunk) {
        const delta = chunk.choices[0]?.delta;
        if (delta.content) {
            console.log(delta.content);
        }
    }
}

There are more features of course, check each library’s README!

Formatting & testing

sudo corepack enable
pnpm install

pnpm -r format:check
pnpm -r lint:check
pnpm -r test

Building

pnpm -r build

This will generate ESM and CJS javascript files in packages/*/dist, eg packages/inference/dist/index.mjs.