Use n8n with MaxonForge
n8n's AI nodes accept an OpenAI credential with a custom base URL. Point that credential at MaxonForge and every workflow run is paid from a prepaid rupee balance, with each call metered separately.
Set it up
- 1
Create an API key
Top up from ₹100 and create a key in Dashboard → API Keys. Give it a name like n8n so its usage is easy to find.
- 2
Create an OpenAI credential
In n8n, go to Credentials → New → OpenAI. Paste the key and set the base URL:
API Key: sk_mf_your_key Base URL: https://www.maxonforge.com/v1
- 3
Use it in an AI node
In an OpenAI Chat Model node (or the OpenAI node), select the credential and enter a model ID from the catalogue. If the model list does not load, type the ID by hand.
MODEL_ID
Check that it works
If a node fails, run the same key and model outside n8n:
curl https://www.maxonforge.com/v1/chat/completions \
-H "Authorization: Bearer sk_mf_your_key" \
-H "Content-Type: application/json" \
-d '{"model":"MODEL_ID","messages":[{"role":"user","content":"Hello"}]}'Replace sk_mf_your_key with your key and MODEL_ID with an ID from the catalogue.
Models you can use with n8n
USD per 1M tokens. Use the model ID exactly as shown. List price is the provider's published rate.
| Model ID | You pay in / out | List in / out |
|---|---|---|
| gpt-6-astra | $1.00 / $5.00 | $10.00 / $50.00 |
| claude-fable-5-1 | $1.00 / $5.00 | $10.00 / $50.00 |
| claude-fable-5 | $1.00 / $5.00 | $10.00 / $50.00 |
| claude-opus-5-5 | $0.400 / $2.00 | $4.00 / $20.00 |
| claude-opus-5 | $0.500 / $2.50 | $5.00 / $25.00 |
| claude-sonnet-5-5 | $0.240 / $1.20 | $2.00 / $10.00 |
Common errors
Credential test fails
The base URL is missing /v1 or still points to OpenAI.
Set it to exactly https://www.maxonforge.com/v1.
401
The key is missing, mistyped or revoked.
Copy the key again from Dashboard → API Keys. MaxonForge website keys start with sk_mf_.
402 insufficient_credits
The prepaid balance cannot cover the request. Nothing was charged.
Top up from ₹100 in Dashboard → Top up, then retry the same call.
404 model_not_found
The model ID is not in the published catalogue.
Copy an exact ID from the model catalogue at /models.
Questions
- How do I keep automation costs under control?
- Give each workflow its own key and restrict it to the models it needs; Dashboard → Usage shows cost per key and per request.
- Does a failed run cost money?
- A request refused for an empty balance is never charged. Requests the model processed are billed at the published rate.
- Self-hosted or n8n cloud?
- Both work. The credential only needs the base URL and key.
Get a key and start from ₹100
Pay with UPI or card. Credit never expires, and every request is metered at the published rate.