OpenRouter API Tester

Read this before you paste a live API key

If none of this sits right for a particular credential, build the request here with the key field empty, press Copy as cURL, and add the key on your own machine. You get the exact request and the secret never leaves your laptop.

How to test an OpenRouter API key

1. Create a key

Keys are created in your OpenRouter account settings. Each one can be given a label and an optional credit limit, which is a ceiling on that key's spending regardless of the account balance.

2. Ask the API about the key

The request pre-filled above is a GET to https://openrouter.ai/api/v1/key. It returns a description of the key that made the call. Do not use the models listing as your key check on this API: /api/v1/models is public and returns 200 with no key at all, so a successful response there proves nothing about your credential.

3. Send the key as a bearer token

The Auth tab is already set to Bearer token; paste the key into the field. It is sent as Authorization: Bearer <key>, as the bearer token guide describes.

4. Send and read the response

A 200 with a data object means the key is valid. The fields inside tell you whether it can actually pay for a request.

Valid key vs invalid key: what you'll see

GET https://openrouter.ai/api/v1/key
Authorization: Bearer <your key>

200 OK → { "data": { "label": "…", "usage": …, "limit": …,
                     "limit_remaining": …, "is_free_tier": …, … } }

401    → { "error": { "code": 401, "message": "…" } }

usage is what the key has spent and limit is its cap, both in US dollars; a limit of null means the key has no cap of its own and is bounded only by the account's credit. limit_remaining is the figure to watch on a capped key, and limit_reset says whether the cap renews on a schedule. is_free_tier marks a free-tier key. The response also breaks usage down by day, week and month.

Notice the error format. error.code is a number that repeats the HTTP status, where OpenAI's format puts a string such as invalid_api_key in that field. A client written for OpenAI that switches on string codes will not match anything here.

Reading errors from a router

With a single-provider API, an error is about your request or about that provider. OpenRouter adds a layer: your request goes to OpenRouter, which chooses a provider for the model and forwards it. An error can come from either hop, and telling them apart decides what you do next.

When the failure is upstream, the error.metadata object often carries the provider's name and its raw error, which is the detail to put in a bug report.

An error inside a 200

HTTP sends the status line before the body. Once a model has started producing output, the status has already gone out as 200, and a failure after that point cannot change it. OpenRouter reports such a failure as an error object in the response body, or as an error event in the stream. A client that checks only response.ok will treat a half-finished generation as a success. Check for a top-level error field, and check each choice's finish_reason, before using the text.

Sending a chat request

POST https://openrouter.ai/api/v1/chat/completions
Authorization: Bearer <your key>
Content-Type: application/json

{
  "model": "<vendor/model-id-from-the-models-list>",
  "messages": [
    { "role": "user", "content": "Say hello." }
  ]
}

Model identifiers on OpenRouter have two parts, a vendor and a model name joined by a slash. Copy one from the GET /api/v1/models response rather than typing it from memory; the id a provider uses on its own API is usually not the id OpenRouter uses for the same model. The catalogue also lists each model's context length and pricing, so that listing, not this page, is where to look for numbers.

To run the call, pick the chat endpoint from the list beside the console, add Content-Type: application/json in the Headers tab and paste the body into the Body tab. This console waits 30 seconds and shows one completed response, so leave streaming off.

The optional attribution headers

HTTP-Referer and X-OpenRouter-Title tell OpenRouter which application a request came from, for its public usage rankings; older examples call the second one X-Title. They are optional and have no effect on authentication. If a request works here without them and fails in your app, they are not the cause.

Using an OpenAI SDK

Set the base URL to https://openrouter.ai/api/v1 and supply the OpenRouter key. The request format is OpenAI's; the model id and the error format are OpenRouter's own.

Provider APIs change — check the official docs

api-tester.com is an independent tool with no affiliation to, partnership with or endorsement from OpenRouter. Nothing here is official documentation. Endpoints, model names, error wording and limits are the provider's to change, and this page can fall behind. For current and authoritative detail, consult the official OpenRouter documentation.

We have deliberately not quoted prices, rate limits or model identifiers. Those change fastest, and OpenRouter's own models endpoint returns them live. What this page describes is the durable shape of the API: how it authenticates, how its errors are structured, and which hop an error came from.

Test other providers and related tools

Frequently asked questions