Mistral Typescript Client

Summary

Mistral AI API: Our Chat Completion and Embeddings APIs specification. Create your account on La Plateforme to get access and read the docs to learn how to use it.

Table of Contents

• Mistral Typescript Client
  • SDK Installation
  • Requirements
  • API Key Setup
  • SDK Example Usage
  • Providers' SDKs
  • Available Resources and Operations
  • Server-sent event streaming
  • File uploads
  • Retries
  • Error Handling
  • Server Selection
  • Custom HTTP Client
  • Authentication
  • Providers Support
  • Standalone functions
  • Debugging
• Development
  • Contributions

SDK Installation

The SDK can be installed with either npm, pnpm, bun or yarn package managers.

NPM

npm add @mistralai/mistralai

PNPM

pnpm add @mistralai/mistralai

Bun

bun add @mistralai/mistralai

Yarn

yarn add @mistralai/mistralai zod

# Note that Yarn does not install peer dependencies automatically. You will need
# to install zod as shown above.

Requirements

For supported JavaScript runtimes, please consult RUNTIMES.md.

API Key Setup

Before you begin, you will need a Mistral AI API key.

1. Get your own Mistral API Key: https://docs.mistral.ai/#api-access
2. Set your Mistral API Key as an environment variable. You only need to do this once.

# set Mistral API Key (using zsh for example)
$ echo 'export MISTRAL_API_KEY=[your_key_here]' >> ~/.zshenv

# reload the environment (or just quit and open a new terminal)
$ source ~/.zshenv

SDK Example Usage

Create Chat Completions

This example shows how to create chat completions.

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.chat.complete({
    model: "mistral-small-latest",
    messages: [
      {
        content:
          "Who is the best French painter? Answer in one short sentence.",
        role: "user",
      },
    ],
  });

  // Handle the result
  console.log(result);
}

run();

Upload a file

This example shows how to upload a file.

import { Mistral } from "@mistralai/mistralai";
import { openAsBlob } from "node:fs";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.files.upload({
    file: await openAsBlob("example.file"),
  });

  // Handle the result
  console.log(result);
}

run();

Create Agents Completions

This example shows how to create agents completions.

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.agents.complete({
    messages: [
      {
        content:
          "Who is the best French painter? Answer in one short sentence.",
        role: "user",
      },
    ],
    agentId: "<value>",
  });

  // Handle the result
  console.log(result);
}

run();

Create Embedding Request

This example shows how to create embedding request.

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.embeddings.create({
    inputs: [
      "Embed this sentence.",
      "As well as this one.",
    ],
    model: "Wrangler",
  });

  // Handle the result
  console.log(result);
}

run();

Providers' SDKs

We have dedicated SDKs for the following providers:

• GCP
• Azure

Available Resources and Operations

agents

• complete - Agents Completion
• stream - Stream Agents completion

batch

batch.jobs

• list - Get Batch Jobs
• create - Create Batch Job
• get - Get Batch Job
• cancel - Cancel Batch Job

chat

• complete - Chat Completion
• stream - Stream chat completion

classifiers

• moderate - Moderations
• moderateChat - Moderations Chat

embeddings

• create - Embeddings

files

• upload - Upload File
• list - List Files
• retrieve - Retrieve File
• delete - Delete File
• download - Download File
• getSignedUrl - Get Signed Url

fim

• complete - Fim Completion
• stream - Stream fim completion

fineTuning

fineTuning.jobs

• list - Get Fine Tuning Jobs
• create - Create Fine Tuning Job
• get - Get Fine Tuning Job
• cancel - Cancel Fine Tuning Job
• start - Start Fine Tuning Job

models

• list - List Models
• retrieve - Retrieve Model
• delete - Delete Model
• update - Update Fine Tuned Model
• archive - Archive Fine Tuned Model
• unarchive - Unarchive Fine Tuned Model

Server-sent event streaming

Server-sent events are used to stream content from certain operations. These operations will expose the stream as an async iterable that can be consumed using a for await...of loop. The loop will terminate when the server no longer has any events to send and closes the underlying connection.

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.chat.stream({
    model: "mistral-small-latest",
    messages: [
      {
        content:
          "Who is the best French painter? Answer in one short sentence.",
        role: "user",
      },
    ],
  });

  for await (const event of result) {
    // Handle the event
    console.log(event);
  }
}

run();

File uploads

Certain SDK methods accept files as part of a multi-part request. It is possible and typically recommended to upload files as a stream rather than reading the entire contents into memory. This avoids excessive memory consumption and potentially crashing with out-of-memory errors when working with very large files. The following example demonstrates how to attach a file stream to a request.

[!TIP]

Depending on your JavaScript runtime, there are convenient utilities that return a handle to a file without reading the entire contents into memory:

• Node.js v20+: Since v20, Node.js comes with a native openAsBlob function in node:fs.
• Bun: The native Bun.file function produces a file handle that can be used for streaming file uploads.
• Browsers: All supported browsers return an instance to a File when reading the value from an <input type="file"> element.
• Node.js v18: A file stream can be created using the fileFrom helper from fetch-blob/from.js.

import { Mistral } from "@mistralai/mistralai";
import { openAsBlob } from "node:fs";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.files.upload({
    file: await openAsBlob("example.file"),
  });

  // Handle the result
  console.log(result);
}

run();

Retries

Some of the endpoints in this SDK support retries. If you use the SDK without any configuration, it will fall back to the default retry strategy provided by the API. However, the default retry strategy can be overridden on a per-operation basis, or across the entire SDK.

To change the default retry strategy for a single API call, simply provide a retryConfig object to the call:

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.models.list({
    retries: {
      strategy: "backoff",
      backoff: {
        initialInterval: 1,
        maxInterval: 50,
        exponent: 1.1,
        maxElapsedTime: 100,
      },
      retryConnectionErrors: false,
    },
  });

  // Handle the result
  console.log(result);
}

run();

If you'd like to override the default retry strategy for all operations that support retries, you can provide a retryConfig at SDK initialization:

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  retryConfig: {
    strategy: "backoff",
    backoff: {
      initialInterval: 1,
      maxInterval: 50,
      exponent: 1.1,
      maxElapsedTime: 100,
    },
    retryConnectionErrors: false,
  },
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.models.list();

  // Handle the result
  console.log(result);
}

run();

Error Handling

All SDK methods return a response object or throw an error. By default, an API error will throw a errors.SDKError.

If a HTTP request fails, an operation my also throw an error from the models/errors/httpclienterrors.ts module:

| HTTP Client Error | Description | | ---------------------------------------------------- | ---------------------------------------------------- | | RequestAbortedError | HTTP request was aborted by the client | | RequestTimeoutError | HTTP request timed out due to an AbortSignal signal | | ConnectionError | HTTP client was unable to make a request to a server | | InvalidRequestError | Any input used to create a request is invalid | | UnexpectedClientError | Unrecognised or unexpected error |

In addition, when custom error responses are specified for an operation, the SDK may throw their associated Error type. You can refer to respective Errors tables in SDK docs for more details on possible error types for each operation. For example, the list method may throw the following errors:

| Error Type | Status Code | Content Type | | -------------------------- | ----------- | ---------------- | | errors.HTTPValidationError | 422 | application/json | | errors.SDKError | 4XX, 5XX | */* |

import { Mistral } from "@mistralai/mistralai";
import {
  HTTPValidationError,
  SDKValidationError,
} from "@mistralai/mistralai/models/errors";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  let result;
  try {
    result = await mistral.models.list();

    // Handle the result
    console.log(result);
  } catch (err) {
    switch (true) {
      case (err instanceof SDKValidationError): {
        // Validation errors can be pretty-printed
        console.error(err.pretty());
        // Raw value may also be inspected
        console.error(err.rawValue);
        return;
      }
      case (err instanceof HTTPValidationError): {
        // Handle err.data$: HTTPValidationErrorData
        console.error(err);
        return;
      }
      default: {
        throw err;
      }
    }
  }
}

run();

Validation errors can also occur when either method arguments or data returned from the server do not match the expected format. The SDKValidationError that is thrown as a result will capture the raw value that failed validation in an attribute called rawValue. Additionally, a pretty() method is available on this error that can be used to log a nicely formatted string since validation errors can list many issues and the plain error string may be difficult read when debugging.

Server Selection

Select Server by Name

You can override the default server globally by passing a server name to the server: keyof typeof ServerList optional parameter when initializing the SDK client instance. The selected server will then be used as the default on the operations that use it. This table lists the names associated with the available servers:

| Name | Server | | ---- | ------------------------ | | eu | https://api.mistral.ai |

Example

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  server: "eu",
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.models.list();

  // Handle the result
  console.log(result);
}

run();

Override Server URL Per-Client

The default server can also be overridden globally by passing a URL to the serverURL: string optional parameter when initializing the SDK client instance. For example:

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  serverURL: "https://api.mistral.ai",
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.models.list();

  // Handle the result
  console.log(result);
}

run();

Custom HTTP Client

The TypeScript SDK makes API calls using an HTTPClient that wraps the native Fetch API. This client is a thin wrapper around fetch and provides the ability to attach hooks around the request lifecycle that can be used to modify the request or handle errors and response.

The HTTPClient constructor takes an optional fetcher argument that can be used to integrate a third-party HTTP client or when writing tests to mock out the HTTP client and feed in fixtures.

The following example shows how to use the "beforeRequest" hook to to add a custom header and a timeout to requests and how to use the "requestError" hook to log errors:

import { Mistral } from "@mistralai/mistralai";
import { HTTPClient } from "@mistralai/mistralai/lib/http";

const httpClient = new HTTPClient({
  // fetcher takes a function that has the same signature as native `fetch`.
  fetcher: (request) => {
    return fetch(request);
  }
});

httpClient.addHook("beforeRequest", (request) => {
  const nextRequest = new Request(request, {
    signal: request.signal || AbortSignal.timeout(5000)
  });

  nextRequest.headers.set("x-custom-header", "custom value");

  return nextRequest;
});

httpClient.addHook("requestError", (error, request) => {
  console.group("Request Error");
  console.log("Reason:", `${error}`);
  console.log("Endpoint:", `${request.method} ${request.url}`);
  console.groupEnd();
});

const sdk = new Mistral({ httpClient });

Authentication

Per-Client Security Schemes

This SDK supports the following security scheme globally:

| Name | Type | Scheme | Environment Variable | | -------- | ---- | ----------- | -------------------- | | apiKey | http | HTTP Bearer | MISTRAL_API_KEY |

To authenticate with the API the apiKey parameter must be set when initializing the SDK client instance. For example:

import { Mistral } from "@mistralai/mistralai";

const mistral = new Mistral({
  apiKey: process.env["MISTRAL_API_KEY"] ?? "",
});

async function run() {
  const result = await mistral.models.list();

  // Handle the result
  console.log(result);
}

run();

Providers Support

We also provide provider specific SDK for:

• GCP
• Azure

Standalone functions

All the methods listed above are available as standalone functions. These functions are ideal for use in applications running in the browser, serverless runtimes or other environments where application bundle size is a primary concern. When using a bundler to build your application, all unused functionality will be either excluded from the final bundle or tree-shaken away.

To read more about standalone functions, check FUNCTIONS.md.

• agentsComplete - Agents Completion
• agentsStream - Stream Agents completion
• batchJobsCancel - Cancel Batch Job
• batchJobsCreate - Create Batch Job
• batchJobsGet - Get Batch Job
• batchJobsList - Get Batch Jobs
• chatComplete - Chat Completion
• chatStream - Stream chat completion
• classifiersModerate - Moderations
• classifiersModerateChat - Moderations Chat
• embeddingsCreate - Embeddings
• filesDelete - Delete File
• filesDownload - Download File
• filesGetSignedUrl - Get Signed Url
• filesList - List Files
• filesRetrieve - Retrieve File
• filesUpload - Upload File
• fimComplete - Fim Completion
• fimStream - Stream fim completion
• fineTuningJobsCancel - Cancel Fine Tuning Job
• fineTuningJobsCreate - Create Fine Tuning Job
• fineTuningJobsGet - Get Fine Tuning Job
• fineTuningJobsList - Get Fine Tuning Jobs
• fineTuningJobsStart - Start Fine Tuning Job
• modelsArchive - Archive Fine Tuned Model
• modelsDelete - Delete Model
• modelsList - List Models
• modelsRetrieve - Retrieve Model
• modelsUnarchive - Unarchive Fine Tuned Model
• modelsUpdate - Update Fine Tuned Model

Debugging

You can setup your SDK to emit debug logs for SDK requests and responses.

You can pass a logger that matches console's interface as an SDK option.

[!WARNING] Beware that debug logging will reveal secrets, like API tokens in headers, in log messages printed to a console or files. It's recommended to use this feature only during local development and not in production.

import { Mistral } from "@mistralai/mistralai";

const sdk = new Mistral({ debugLogger: console });

You can also enable a default debug logger by setting an environment variable MISTRAL_DEBUG to true.

Development

Contributions

While we value open-source contributions to this SDK, this library is generated programmatically. Any manual changes added to internal files will be overwritten on the next generation. We look forward to hearing your feedback. Feel free to open a PR or an issue with a proof of concept and we'll do our best to include it in a future release.