Inference API always returns error: Invalid token

[greyblue9]

Sorry if this is not the best place to post this issue. I am having an issue with the inference API suddenly, after it has worked perfectly for months:

After seeing '{ "error": "invalid token" }' coming back in response to queries, I created a new API token (old one was not showing up, abd changed the header from Authorization: Bearer api_XXX to Authorization: Bearer hf_YYY as outlined in the docs, but am still facing the same error.

Any idea what could be the issue? I am so hoping this could be fixed soon.

[Narsil]

Hi @greyblue9 ,

Are you sure you're correctly passing the token ?
We recently changed rules, to be less silent about incorrect tokens especially when formatting was forgotten.
We used to simply pass request as unauthenticated but it caused issues for some users not realizing they had made simple typos and couldn't access their plan features

f"Bearer {API_TOKEN}"  #works
# vs
"Bearer {API_TOKEN}"  # doesn't work, no formatting, `"{API_TOKEN}"` is actually invalid.

Both old and new tokens should work the same way by the way.

Sorry about this issue.

[greyblue9]

Hi @greyblue9 ,
Both old and new tokens should work the same way by the way.

Thank you so much for your response. I actually created a new account with a new (write) token, and that worked 🤔

What I was using previously, the token was much longer (about 3x longer and had no api_ or hf_ prefix on it, and I was using f"Bearer: api_{token}")

Maybe the : was the issue?

[Narsil]

Probably, we really need to get Bearer {API_TOKEN} !

Glad you could work it out.

[Narsil]

I will close this, don ´t hesitate to open again if you think it's incorrect.