Add an AI key

exelens calls models with your keys (bring your own). You pay your provider directly. We never add an AI charge on our invoice. Keys are configured once in Settings and then mostly left alone — that is why they live under the avatar, not in Data.

1. 1

   Open Settings

   Click your avatar (top right) → Settings → AI keys and usage.

2. 2

   Add key

   Click Add key. Pick the provider by its logo: Anthropic, OpenAI, Google Gemini, xAI, Azure OpenAI, Amazon Bedrock, or an OpenAI-compatible endpoint. Local Docker also offers a scripted test model with no key.

3. 3

   Paste the secret

   Name the key so people recognise it later (“Company Anthropic”). Paste the API key, or for Bedrock the role ARN, external ID and region. Optionally mark it Only for me so others cannot use it.

4. 4

   Save and check

   exelens makes one small request to the provider. If it succeeds, the key is encrypted and never shown again. Use Update models any time the provider adds models.

5. 5

   Set a budget

   Open Budget on the key. Set a monthly cap and an optional per-person cap in USD. Leave “Stop new questions when the cap is reached” on if you want a hard stop.

Several keys

You can register as many keys as you need — different providers, a cheaper model for daily boards, a stronger model for hard questions, a personal key beside a company key. In Build, the model picker lists every usable key. Usage (spend, tokens, requests, daily chart) is on the same Settings page.