Your first XEDOC API request

1. Create an account

   Register with a username and password, or continue with GitHub or Discord. Password registration asks you to repeat the password and solve the verification image.

2. Add token credit

   Open Overview and choose Top up tokens. Select an amount, optionally apply a promo code, and continue to the secure NOWPayments crypto checkout. Tokens are credited after the payment reaches its final successful status.

3. Create a key

   Open Dashboard → API keys, choose a name and create the key. Copy it when shown; the full secret is displayed only once.

4. Check access

   API access is enabled automatically when you register. If an administrator disables access, generation returns 403; insufficient token balance returns 402. Check your account status and token balance in Overview.

Listing models checks authentication without generating content.

Invoke-RestMethod -Uri "https://api.xedoc.top/v1/models" -Headers @{ Authorization = "Bearer $env:XEDOC_API_KEY" }

curl "https://api.xedoc.top/v1/models" \
  -H "Authorization: Bearer $XEDOC_API_KEY"

curl "https://api.xedoc.top/v1/models" \
  -H "Authorization: Bearer $XEDOC_API_KEY"

The response contains a data array with the available canonical model IDs. Use those IDs exactly as returned.

$body = @{
  model = "gpt-5.6-terra"
  input = "Reply with a short hello."
} | ConvertTo-Json

Invoke-RestMethod -Uri "https://api.xedoc.top/v1/responses" -Method Post -Headers @{ Authorization = "Bearer $env:XEDOC_API_KEY" } -ContentType "application/json" -Body $body

curl "https://api.xedoc.top/v1/responses" \
  -H "Authorization: Bearer $XEDOC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.6-terra","input":"Reply with a short hello."}'

curl "https://api.xedoc.top/v1/responses" \
  -H "Authorization: Bearer $XEDOC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.6-terra","input":"Reply with a short hello."}'

Generation deducts model-weighted tokens from your balance. Open Requests to inspect raw token usage and the actual charge separately. If the model is unavailable, select another ID from your current model list. Use the HTTP error reference to diagnose a rejected request.