Jev API key guide · updated October 4, 2026
How to get a Jev API key
A Jev API key on Jev AI takes about a minute: sign in, create a key on the API page and store it on your server. Eligible new accounts start with 20 welcome credits, enough to verify the key and send your first decision requests.
Which Jev API key do you need?
Two separate services issue keys for the Jev model. Decide which one you are calling before you create anything.
| Jev AI key (this site) | TypeSafe key (official) | |
|---|---|---|
| Where you create it | The Jev API page on jev-ai.pro | TypeSafe’s own console |
| Endpoint it authenticates | https://jev-ai.pro/api/v1/systemone | TypeSafe’s official API |
| Balance it spends | Your Jev AI credits and tokens | Your TypeSafe account |
| Works with the TypeSafe SDK | Yes, with baseURL set to https://jev-ai.pro/api | Yes, with the SDK defaults |
| Interchangeable | No. A Jev AI key is rejected by TypeSafe | No. A TypeSafe key is rejected here |
TypeSafe develops Jev and runs its own console and API. Jev AI is an independently operated playground and API with a TypeSafe-compatible endpoint. For the official service, start from the TypeSafe documentation. The rest of this guide covers the Jev AI key.
Create your key in five steps
No card is needed to sign in or to create a key. Creating a key does not run the model.
- 01
Sign in to Jev AI
Use the sign-in button in the header. There is no password to set. Your keys, usage and balance belong to this account.
- 02
Open “Your Jev API keys”
Go to the Jev API page. Give the key a name of up to 40 characters, such as the app or environment that will use it, then select Create key.
- 03
Copy the secret now
The full key, beginning with
sk_jev-ai_, is displayed once. Afterwards the list shows only a prefix, the last four characters and when the key was last used. - 04
Store it as JEV_AI_API_KEY
Add it to your server environment and your deployment secrets. Keep it out of browser code, logs, commits and agent chats.
- 05
Verify, then call
List the models with the key to confirm it works, then send one small decision request. Both commands are below.
Test your Jev API key
Check authentication first. It costs nothing and separates a key problem from a request problem.
1. Confirm the key is accepted
# Lists the models your key can call. Runs no inference and uses no balance.
curl --fail-with-body https://jev-ai.pro/api/v1/models \
-H "Authorization: Bearer $JEV_AI_API_KEY"A models array means the key and the host are right. A 401 means the key was not accepted.
2. Send your first decision
curl --fail-with-body https://jev-ai.pro/api/v1/systemone \
-H "Authorization: Bearer $JEV_AI_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"jev-latest","state":"My payment failed. Please help.","questions":{"urgent":{"type":"noul","instructions":"Does this message need urgent support?"}}}'Read answers.urgent.noul in the response: the probability of yes, from 0 to 1. This call is metered. The Jev API page has a version that reads the key from a hidden prompt, plus a button that confirms your first successful call.
What a new key costs to use
The key is free. Requests are paid from your Jev AI balance.
- Welcome credits: eligible new accounts receive 20 credits. They work in the playground and as the fallback for API calls.
- API billing order: an API call uses purchased input tokens first. If tokens cannot cover it, the whole call is paid with the model’s credit charge instead. A request never charges both.
- Output tokens: free. Only input tokens are metered.
- Failed calls: an upstream failure or timeout returns the reserved balance.
See what is free on Jev AI for the per-model credit cost, or API pricing for token plans.
Using the key with the TypeSafe SDK
Changing only the key is the most common setup mistake.
The @typesafe-ai/sdk package is published by TypeSafe and defaults to TypeSafe’s service. To send requests to Jev AI, set baseURL: 'https://jev-ai.pro/api' together with your Jev AI key. The SDK appends /v1/systemone itself, so leave /v1 out of the base URL.
The SDK setup section of the docs has a complete example and a connection checklist. Prefer plain HTTP? The Jev tutorial walks through a full request in cURL and JavaScript.
When the key does not work
Fix configuration errors before sending more decision requests.
| Status | Likely cause | What to do |
|---|---|---|
401 | The key is missing, mistyped or revoked, or a Jev AI key was sent to another host. | Check the Bearer header and the destination. With the TypeSafe SDK, set baseURL as well as the key. |
402 | Neither your tokens nor your credits can cover the request. | Check your balance on the Jev API page. A 402 is not charged. |
404 | The path is wrong, usually a missing /api or a repeated /v1 in the base URL. | The final request must reach https://jev-ai.pro/api/v1/systemone. |
429 | More than 1,000 API requests per minute on the account, or Jev is busy. | Back off and honor Retry-After when it is present. |
The full list of responses is in errors and retries. To rotate a key, create the replacement first, deploy it, then revoke the old key on the Jev API page.
Jev API key FAQ
How do I get a Jev API key?
Sign in to Jev AI, open the Jev API page, name a key under “Your Jev API keys” and select Create key. Copy the secret when it appears, because it is shown only once, and store it on your server as JEV_AI_API_KEY.
Is a Jev API key free?
Creating a key is free and does not run the model. Eligible new accounts receive 20 welcome credits. API calls use purchased input tokens first; when there are none, a call is paid with credits instead, so the welcome credits cover your first requests.
Can I use my TypeSafe API key on Jev AI?
No. Jev AI is an independent service with its own accounts, keys and balances. A key from TypeSafe authenticates TypeSafe’s official API, and a Jev AI key authenticates the endpoint on jev-ai.pro. Pair each key with its own endpoint.
I lost my key. Can I see it again?
No. The full secret is shown once, at creation. The key list only displays a prefix and the last four characters. Create a new key, update your server, then revoke the old one.
What does a Jev AI key look like?
A Jev AI key is a single string that begins with sk_jev-ai_. Send it in the Authorization header as a Bearer token.
Which models can one key call?
One key covers every model that GET /api/v1/models lists for it, including the Jev aliases jev-latest, jev-preview and jev-1.13.0. Call that endpoint before you send a request; it runs no inference.
Is there a rate limit on the Jev API?
Yes. Jev AI allows 1,000 API requests per minute per account, and the live web context route has a separate limit of 120 requests per minute. A 429 response may include Retry-After.
Where should I store the key?
In a server-side environment variable or your deployment’s secret store. Never put it in browser code, logs, commits or a chat with a coding agent. If a key is exposed, revoke it on the Jev API page and create a new one.
Ready to create your key?
Sign in, name a key and copy it once. Then verify it with the models endpoint.
Jev AI is independently operated and is not TypeSafe’s official site. Limits and billing rules describe Jev AI as of October 4, 2026.