Skip to main content
Question

API Failed - Hugging Face (AI: Serverless Inference)

  • March 6, 2026
  • 1 reply
  • 16 views
K
Ground Control | Tier 1
Forum|alt.badge.img

I am trying to use AI: Serverless Inference in Automation Anywhere 360 with Hugging Face.

Flow:

  • AI: Authentication (Hugging Face token)

  • AI: Serverless Inference (tested models like gpt2, google/flan-t5-small)

  • Message Box to display response

However, the bot always goes to the Catch block. When I printed the error message, I got:

Status: 410
https://api-inference.huggingface.co is no longer supported.
Please use https://router.huggingface.co instead.

It looks like Automation Anywhere is calling the deprecated Hugging Face API endpoint.

Has anyone faced this issue?
How can we configure Automation Anywhere to use the new endpoint https://router.huggingface.co?

1 reply

Aaron.Gleason
Automation Anywhere Team
Forum|alt.badge.img+14
  • Aaron.Gleason
  • Automation Anywhere Team
  • March 6, 2026

​@krishna_kr The only fix I can think of is to create a custom model connection. To do this, you must use Automation Anywhere APIs. There is no user interface at the moment that allows creation of custom models.

https://{yourcontrolroom}/swagger/ui/?url=/swagger/api/v1/ai-agent-studio-api.yaml

I made a video last year on the topic:

https://upskill.automationanywhere.com/getting-started-with-custom-generative-ai-model-connections

In the meantime, I would recommend starting a support ticket by emailing support@automationanywhere.com and mentioning the domain change causing the AI package to need updating.

Terms & ConditionsCookie settingsAccessibility statement