The Venice ai API key generated on Agent Zero's Dashboard (website) FAILS TO AUTHENTICATE!

[sumer-man]

I've got A0T tokens staked and the agent-zero's API Dashboard says that I do have my daily free AI inference credits allowance worth ~$0.70 atm.

I have configured my running agent-zero docker container as follows:

• added the API key to all the relevant places in A0 > dashboard > Settings
• set the provider to Venice.ai
• chose the right model (official model name copied from the venice.ai website, as recommended)

but getting an Authentication error both in the A0 chat and in the Docker (detailed error is below).
Then I curl'ed the venice.ai API directly with:

curl https://api.venice.ai/api/v1/chat/completions \
-H "Authorization: Bearer $A0_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "venice-uncensored",
"messages": [{"role": "user", "content": "Hello World!"}]
}'

and got:
{"error":"Authentication failed"}

Note: The A0_API_KEY is properly exported in the current terminal/bash session.

Then I've also tested my A0 (and curl) with a token generated straight from Venice.ai site, and got:
litellm.exceptions.APIError: litellm.APIError: APIError: OpenAIException - Error code: 402 - {'error': 'Insufficient USD or Diem balance to complete request'}

Meaning both, the Venice.ai Auth service and API work fine, and the Venice.ai API integration within the latest ver. of A0 (docker image) is correct.
Therefore it must be the token that is somewhat invalid.

The token generated on the A0 Dashboard looks like:
sk-a0-5fVAVOznKSaacfInoQqIRRbNDIfMs99AhH

whereas the one generated on Venice.ai site looks like this:
VENICE-INFERENCE-KEY-uPOVhX2mQmrtg9J0SxGS7OWQjOJjwIbjO7EyMMwuxb

What's going on?
Best Regards

P.S. The detailed A0/Docker Auth error:

litellm.exceptions.AuthenticationError: litellm.AuthenticationError: AuthenticationError: OpenAIException - Error code: 401 - {'error': 'Authentication failed'}

Traceback (most recent call last):
Traceback (most recent call last):
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/llms/openai/[openai.py](https://openai.py/)", line 991, in async_streaming
    headers, response = await self.make_openai_chat_completion_request(
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/litellm_core_utils/logging_[utils.py](https://utils.py/)", line 190, in async_wrapper
    result = await func(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/llms/openai/[openai.py](https://openai.py/)", line 454, in make_openai_chat_completion_request
    raise e
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/llms/openai/[openai.py](https://openai.py/)", line 436, in make_openai_chat_completion_request
    await openai_[aclient.chat](https://aclient.chat/).completions.with_raw_response.create(
  File "/opt/venv-a0/lib/python3.12/site-packages/openai/_legacy_[response.py](https://response.py/)", line 381, in wrapped
    return cast(LegacyAPIResponse[R], await func(*args, **kwargs))
                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/openai/resources/chat/completions/[completions.py](https://completions.py/)", line 2589, in create
    return await self._post(
           ^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/openai/_base_[client.py](https://client.py/)", line 1794, in post
    return await self.request(cast_to, opts, stream=stream, stream_cls=stream_cls)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/openai/_base_[client.py](https://client.py/)", line 1594, in request
    raise self._make_status_error_from_response(err.response) from None
openai.AuthenticationError: Error code: 401 - {'error': 'Authentication failed'}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/[main.py](https://main.py/)", line 598, in acompletion
    response = await init_response
               ^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/llms/openai/[openai.py](https://openai.py/)", line 1041, in async_streaming
    raise OpenAIError(
litellm.llms.openai.common_utils.OpenAIError: Error code: 401 - {'error': 'Authentication failed'}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/a0/[agent.py](https://agent.py/)", line 454, in monologue
    agent_response, _reasoning = await [self.call](https://self.call/)_chat_model(
                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/a0/[agent.py](https://agent.py/)", line 808, in call_chat_model
    response, reasoning = await model.unified_call(
                          ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/a0/[models.py](https://models.py/)", line 502, in unified_call
    _completion = await acompletion(
                  ^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/[utils.py](https://utils.py/)", line 1638, in wrapper_async
    raise e
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/[utils.py](https://utils.py/)", line 1484, in wrapper_async
    result = await original_function(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/[main.py](https://main.py/)", line 617, in acompletion
    raise exception_type(
          ^^^^^^^^^^^^^^^
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/litellm_core_utils/exception_mapping_[utils.py](https://utils.py/)", line 2323, in exception_type
    raise e
  File "/opt/venv-a0/lib/python3.12/site-packages/litellm/litellm_core_utils/exception_mapping_[utils.py](https://utils.py/)", line 477, in exception_type
    raise AuthenticationError(

[frdel]

Set provider to Agent Zero, should be the first in the list.

[frdel]

Also revoke your api key and generate a new one since you posted it here. Keep your api keys private.

[sumer-man]

Cheers!
I managed to figure it out eventually.
The thing is that on the agent zero dashboard it still says that you should use Venice AI as your model provider.
It literally says:
"
How to use
Stake some A0T tokens to unlock free daily API credits. [See stakes]
Generate your API key in the LLM API key card on this page.
For Chat, Utility, or Browser model, select the Agent Zero Venice.ai model provider.
Go to the External Services tab, open API keys, and paste your API key into the Agent Zero Venice.ai field.
Select one of the Venice text models: https://docs.venice.ai/models/text
Enjoy your free daily quota.
"
So I did, and it didn't work.
Eventually, with the help of a different LLM provider which worked within agent-zero, we figured out that the provider should be set to the "Agent Zero API", now available in the settings as chat model provider.
Please update your docs and the website then close this issue.
So confusing atm.
BTW, no worries, these are not the real API keys that I used. They're just examples of what they look like :-)