This library provides convenient access to the OpenAI REST API from TypeScript or JavaScript.
It is generated from ourOpenAPI specificationwithStainless.
To learn how to use the OpenAI API, check out ourAPI ReferenceandDocumentation.
npm install openai
You can import in Deno via:
importOpenAIfrom'https://deno.land/x/[email protected]/mod.ts';
The full API of this library can be found inapi.md filealong with manycode examples.The code below shows how to get started using the chat completions API.
importOpenAIfrom'openai';
constclient=newOpenAI({
apiKey:process.env['OPENAI_API_KEY'],// This is the default and can be omitted
});
asyncfunctionmain(){
constchatCompletion=awaitclient.chat.completions.create({
messages:[{role:'user',content:'Say this is a test'}],
model:'gpt-3.5-turbo',
});
}
main();
We provide support for streaming responses using Server Sent Events (SSE).
importOpenAIfrom'openai';
constclient=newOpenAI();
asyncfunctionmain(){
conststream=awaitclient.chat.completions.create({
model:'gpt-4',
messages:[{role:'user',content:'Say this is a test'}],
stream:true,
});
forawait(constchunkofstream){
process.stdout.write(chunk.choices[0]?.delta?.content||'');
}
}
main();
If you need to cancel a stream, you canbreak
from the loop
or callstream.controller.abort()
.
This library includes TypeScript definitions for all request params and response fields. You may import and use them like so:
importOpenAIfrom'openai';
constclient=newOpenAI({
apiKey:process.env['OPENAI_API_KEY'],// This is the default and can be omitted
});
asyncfunctionmain(){
constparams:OpenAI.Chat.ChatCompletionCreateParams={
messages:[{role:'user',content:'Say this is a test'}],
model:'gpt-3.5-turbo',
};
constchatCompletion:OpenAI.Chat.ChatCompletion=awaitclient.chat.completions.create(params);
}
main();
Documentation for each method, request param, and response field are available in docstrings and will appear on hover in most modern editors.
Important
Previous versions of this SDK used aConfiguration
class. See thev3 to v4 migration guide.
When interacting with the API some actions such as starting a Run and adding files to vector stores are asynchronous and take time to complete. The SDK includes helper functions which will poll the status until it reaches a terminal state and then return the resulting object. If an API method results in an action which could benefit from polling there will be a corresponding version of the method ending in 'AndPoll'.
For instance to create a Run and poll until it reaches a terminal state you can run:
construn=awaitopenai.beta.threads.runs.createAndPoll(thread.id,{
assistant_id:assistantId,
});
More information on the lifecycle of a Run can be found in theRun Lifecycle Documentation
When creating and interacting with vector stores, you can use the polling helpers to monitor the status of operations. For convenience, we also provide a bulk upload helper to allow you to simultaneously upload several files at once.
constfileList=[
createReadStream('/home/data/example.pdf'),
...
];
constbatch=awaitopenai.vectorStores.fileBatches.uploadAndPoll(vectorStore.id,fileList);
The SDK also includes helpers to process streams and handle the incoming events.
construn=openai.beta.threads.runs
.stream(thread.id,{
assistant_id:assistant.id,
})
.on('textCreated',(text)=>process.stdout.write('\nassistant > '))
.on('textDelta',(textDelta,snapshot)=>process.stdout.write(textDelta.value))
.on('toolCallCreated',(toolCall)=>process.stdout.write(`\nassistant >${toolCall.type}\n\n`))
.on('toolCallDelta',(toolCallDelta,snapshot)=>{
if(toolCallDelta.type==='code_interpreter'){
if(toolCallDelta.code_interpreter.input){
process.stdout.write(toolCallDelta.code_interpreter.input);
}
if(toolCallDelta.code_interpreter.outputs){
process.stdout.write('\noutput >\n');
toolCallDelta.code_interpreter.outputs.forEach((output)=>{
if(output.type==='logs'){
process.stdout.write(`\n${output.logs}\n`);
}
});
}
}
});
More information on streaming helpers can be found in the dedicated documentation:helpers.md
This library provides several conveniences for streaming chat completions, for example:
importOpenAIfrom'openai';
constopenai=newOpenAI();
asyncfunctionmain(){
conststream=awaitopenai.beta.chat.completions.stream({
model:'gpt-4',
messages:[{role:'user',content:'Say this is a test'}],
stream:true,
});
stream.on('content',(delta,snapshot)=>{
process.stdout.write(delta);
});
// or, equivalently:
forawait(constchunkofstream){
process.stdout.write(chunk.choices[0]?.delta?.content||'');
}
constchatCompletion=awaitstream.finalChatCompletion();
console.log(chatCompletion);// {id: "…", choices: […],…}
}
main();
Streaming withopenai.beta.chat pletions.stream({…})
exposes
various helpers for your convenienceincluding event handlers and promises.
Alternatively, you can useopenai.chat pletions.create({ stream: true,… })
which only returns an async iterable of the chunks in the stream and thus uses less memory
(it does not build up a final chat completion object for you).
If you need to cancel a stream, you canbreak
from afor await
loop or callstream.abort()
.
We provide theopenai.beta.chat pletions.runTools({…})
convenience helper for using function tool calls with the/chat/completions
endpoint
which automatically call the JavaScript functions you provide
and sends their results back to the/chat/completions
endpoint,
looping as long as the model requests tool calls.
If you pass aparse
function, it will automatically parse thearguments
for you
and returns any parsing errors to the model to attempt auto-recovery.
Otherwise, the args will be passed to the function you provide as a string.
If you passtool_choice: {function: {name:…}}
instead ofauto
,
it returns immediately after calling that function (and only loops to auto-recover parsing errors).
importOpenAIfrom'openai';
constclient=newOpenAI();
asyncfunctionmain(){
construnner=client.beta.chat.completions
.runTools({
model:'gpt-3.5-turbo',
messages:[{role:'user',content:'How is the weather this week?'}],
tools:[
{
type:'function',
function:{
function:getCurrentLocation,
parameters:{type:'object',properties:{}},
},
},
{
type:'function',
function:{
function:getWeather,
parse:JSON.parse,// or use a validation library like zod for typesafe parsing.
parameters:{
type:'object',
properties:{
location:{type:'string'},
},
},
},
},
],
})
.on('message',(message)=>console.log(message));
constfinalContent=awaitrunner.finalContent();
console.log();
console.log('Final content:',finalContent);
}
asyncfunctiongetCurrentLocation(){
return'Boston';// Simulate lookup
}
asyncfunctiongetWeather(args:{location:string}){
const{location}=args;
//… do lookup…
return{temperature,precipitation};
}
main();
// {role: "user", content: "How's the weather this week?" }
// {role: "assistant", tool_calls: [{type: "function", function: {name: "getCurrentLocation", arguments: "{}" }, id: "123" }
// {role: "tool", name: "getCurrentLocation", content: "Boston", tool_call_id: "123" }
// {role: "assistant", tool_calls: [{type: "function", function: {name: "getWeather", arguments: '{ "location": "Boston" }'}, id: "1234" }]}
// {role: "tool", name: "getWeather", content: '{ "temperature": "50degF", "preciptation": "high" }', tool_call_id: "1234" }
// {role: "assistant", content: "It's looking cold and rainy - you might want to wear a jacket!" }
//
// Final content: "It's looking cold and rainy - you might want to wear a jacket!"
Like with.stream()
,we provide a variety ofhelpers and events.
Note thatrunFunctions
was previously available as well, but has been deprecated in favor ofrunTools
.
Read more about various examples such as with integrating withzod, next.js,andproxying a stream to the browser.
Request parameters that correspond to file uploads can be passed in many different forms:
File
(or an object with the same structure)- a
fetch
Response
(or an object with the same structure) - an
fs.ReadStream
- the return value of our
toFile
helper
importfsfrom'fs';
importfetchfrom'node-fetch';
importOpenAI,{toFile}from'openai';
constclient=newOpenAI();
// If you have access to Node `fs` we recommend using `fs.createReadStream()`:
awaitclient.files.create({file:fs.createReadStream('input.jsonl'),purpose:'fine-tune'});
// Or if you have the web `File` API you can pass a `File` instance:
awaitclient.files.create({file:newFile(['my bytes'],'input.jsonl'),purpose:'fine-tune'});
// You can also pass a `fetch` `Response`:
awaitclient.files.create({file:awaitfetch('https://somesite/input.jsonl'),purpose:'fine-tune'});
// Finally, if none of the above are convenient, you can use our `toFile` helper:
awaitclient.files.create({
file:awaittoFile(Buffer.from('my bytes'),'input.jsonl'),
purpose:'fine-tune',
});
awaitclient.files.create({
file:awaittoFile(newUint8Array([0,1,2]),'input.jsonl'),
purpose:'fine-tune',
});
When the library is unable to connect to the API,
or if the API returns a non-success status code (i.e., 4xx or 5xx response),
a subclass ofAPIError
will be thrown:
asyncfunctionmain(){
constjob=awaitclient.fineTuning.jobs
.create({model:'gpt-3.5-turbo',training_file:'file-abc123'})
.catch(async(err)=>{
if(errinstanceofOpenAI.APIError){
console.log(err.status);// 400
console.log(err.name);// BadRequestError
console.log(err.headers);// {server: 'nginx',...}
}else{
throwerr;
}
});
}
main();
Error codes are as followed:
Status Code | Error Type |
---|---|
400 | BadRequestError |
401 | AuthenticationError |
403 | PermissionDeniedError |
404 | NotFoundError |
422 | UnprocessableEntityError |
429 | RateLimitError |
>=500 | InternalServerError |
N/A | APIConnectionError |
To use this library withAzure OpenAI,use theAzureOpenAI
class instead of theOpenAI
class.
Important
The Azure API shape slightly differs from the core API shape which means that the static types for responses / params won't always be correct.
import{AzureOpenAI}from'openai';
import{getBearerTokenProvider,DefaultAzureCredential}from'@azure/identity';
constcredential=newDefaultAzureCredential();
constscope='https://cognitiveservices.azure /.default';
constazureADTokenProvider=getBearerTokenProvider(credential,scope);
constopenai=newAzureOpenAI({azureADTokenProvider});
constresult=awaitopenai.chat.completions.create({
model:'gpt-4-1106-preview',
messages:[{role:'user',content:'Say hello!'}],
});
console.log(result.choices[0]!.message?.content);
Certain errors will be automatically retried 2 times by default, with a short exponential backoff. Connection errors (for example, due to a network connectivity problem), 408 Request Timeout, 409 Conflict, 429 Rate Limit, and >=500 Internal errors will all be retried by default.
You can use themaxRetries
option to configure or disable this:
// Configure the default for all requests:
constclient=newOpenAI({
maxRetries:0,// default is 2
});
// Or, configure per-request:
awaitclient.chat.completions.create({messages:[{role:'user',content:'How can I get the name of the current day in Node.js?'}],model:'gpt-3.5-turbo'},{
maxRetries:5,
});
Requests time out after 10 minutes by default. You can configure this with atimeout
option:
// Configure the default for all requests:
constclient=newOpenAI({
timeout:20*1000,// 20 seconds (default is 10 minutes)
});
// Override per-request:
awaitclient.chat.completions.create({messages:[{role:'user',content:'How can I list all files in a directory using Python?'}],model:'gpt-3.5-turbo'},{
timeout:5*1000,
});
On timeout, anAPIConnectionTimeoutError
is thrown.
Note that requests which time out will beretried twice by default.
List methods in the OpenAI API are paginated.
You can usefor await… of
syntax to iterate through items across all pages:
asyncfunctionfetchAllFineTuningJobs(params){
constallFineTuningJobs=[];
// Automatically fetches more pages as needed.
forawait(constfineTuningJobofclient.fineTuning.jobs.list({limit:20})){
allFineTuningJobs.push(fineTuningJob);
}
returnallFineTuningJobs;
}
Alternatively, you can make request a single page at a time:
letpage=awaitclient.fineTuning.jobs.list({limit:20});
for(constfineTuningJobofpage.data){
console.log(fineTuningJob);
}
// Convenience methods are provided for manually paginating:
while(page.hasNextPage()){
page=page.getNextPage();
//...
}
The "raw"Response
returned byfetch()
can be accessed through the.asResponse()
method on theAPIPromise
type that all methods return.
You can also use the.withResponse()
method to get the rawResponse
along with the parsed data.
constclient=newOpenAI();
constresponse=awaitclient.chat.completions
.create({messages:[{role:'user',content:'Say this is a test'}],model:'gpt-3.5-turbo'})
.asResponse();
console.log(response.headers.get('X-My-Header'));
console.log(response.statusText);// access the underlying Response object
const{data:chatCompletion,response:raw}=awaitclient.chat.completions
.create({messages:[{role:'user',content:'Say this is a test'}],model:'gpt-3.5-turbo'})
.withResponse();
console.log(raw.headers.get('X-My-Header'));
console.log(chatCompletion);
This library is typed for convenient access to the documented API. If you need to access undocumented endpoints, params, or response properties, the library can still be used.
To make requests to undocumented endpoints, you can useclient.get
,client.post
,and other HTTP verbs.
Options on the client, such as retries, will be respected when making these requests.
awaitclient.post('/some/path',{
body:{some_prop:'foo'},
query:{some_query_arg:'bar'},
});
To make requests using undocumented parameters, you may use// @ts-expect-error
on the undocumented
parameter. This library doesn't validate at runtime that the request matches the type, so any extra values you
send will be sent as-is.
client.foo.create({
foo:'my_param',
bar:12,
// @ts-expect-error baz is not yet public
baz:'undocumented option',
});
For requests with theGET
verb, any extra params will be in the query, all other requests will send the
extra param in the body.
If you want to explicitly send an extra argument, you can do so with thequery
,body
,andheaders
request
options.
To access undocumented response properties, you may access the response object with// @ts-expect-error
on
the response object, or cast the response object to the requisite type. Like the request params, we do not
validate or strip extra properties from the response from the API.
By default, this library usesnode-fetch
in Node, and expects a globalfetch
function in other environments.
If you would prefer to use a global, web-standards-compliantfetch
function even in a Node environment,
(for example, if you are running Node with--experimental-fetch
or using NextJS which polyfills withundici
),
add the following import before your first importfrom "OpenAI"
:
// Tell TypeScript and the package to use the global web fetch instead of node-fetch.
// Note, despite the name, this does not add any polyfills, but expects them to be provided if needed.
import'openai/shims/web';
importOpenAIfrom'openai';
To do the inverse, addimport "openai/shims/node"
(which does import polyfills).
This can also be useful if you are getting the wrong TypeScript types forResponse
(more details).
You may also provide a customfetch
function when instantiating the client,
which can be used to inspect or alter theRequest
orResponse
before/after each request:
import{fetch}from'undici';// as one example
importOpenAIfrom'openai';
constclient=newOpenAI({
fetch:async(url:RequestInfo,init?:RequestInit):Promise<Response>=>{
console.log('About to make a request',url,init);
constresponse=awaitfetch(url,init);
console.log('Got response',response);
returnresponse;
},
});
Note that if given aDEBUG=true
environment variable, this library will log all requests and responses automatically.
This is intended for debugging purposes only and may change in the future without notice.
By default, this library uses a stable agent for all http/https requests to reuse TCP connections, eliminating many TCP & TLS handshakes and shaving around 100ms off most requests.
If you would like to disable or customize this behavior, for example to use the API behind a proxy, you can pass anhttpAgent
which is used for all requests (be they http or https), for example:
importhttpfrom'http';
import{HttpsProxyAgent}from'https-proxy-agent';
// Configure the default for all requests:
constclient=newOpenAI({
httpAgent:newHttpsProxyAgent(process.env.PROXY_URL),
});
// Override per-request:
awaitclient.models.list({
httpAgent:newhttp.Agent({keepAlive:false}),
});
This package generally followsSemVerconventions, though certain backwards-incompatible changes may be released as minor versions:
- Changes that only affect static types, without breaking runtime behavior.
- Changes to library internals which are technically public but not intended or documented for external use.(Please open a GitHub issue to let us know if you are relying on such internals).
- Changes that we do not expect to impact the vast majority of users in practice.
We take backwards-compatibility seriously and work hard to ensure you can rely on a smooth upgrade experience.
We are keen for your feedback; please open anissuewith questions, bugs, or suggestions.
TypeScript >= 4.5 is supported.
The following runtimes are supported:
-
Node.js 18 LTS or later (non-EOL) versions.
-
Deno v1.28.0 or higher, using
import OpenAI from "npm:openai"
. -
Bun 1.0 or later.
-
Cloudflare Workers.
-
Vercel Edge Runtime.
-
Jest 28 or greater with the
"node"
environment ("jsdom"
is not supported at this time). -
Nitro v2.6 or greater.
-
Web browsers: disabled by default to avoid exposing your secret API credentials. Enable browser support by explicitly setting
dangerouslyAllowBrowser
to true'.More explanation
Enabling the
dangerouslyAllowBrowser
option can be dangerous because it exposes your secret API credentials in the client-side code. Web browsers are inherently less secure than server environments, any user with access to the browser can potentially inspect, extract, and misuse these credentials. This could lead to unauthorized access using your credentials and potentially compromise sensitive data or functionality.In certain scenarios where enabling browser support might not pose significant risks:
- Internal Tools: If the application is used solely within a controlled internal environment where the users are trusted, the risk of credential exposure can be mitigated.
- Public APIs with Limited Scope: If your API has very limited scope and the exposed credentials do not grant access to sensitive data or critical operations, the potential impact of exposure is reduced.
- Development or debugging purpose: Enabling this feature temporarily might be acceptable, provided the credentials are short-lived, aren't also used in production environments, or are frequently rotated.
Note that React Native is not supported at this time.
If you are interested in other runtime environments, please open or upvote an issue on GitHub.