[GH-ISSUE #30] model authentication #13

New Issue

2026-06-06T22:07:58-04:00

yindo commented

2026-06-06 22:07:58 -04:00

Originally created by @JacksonLvX on GitHub (Apr 8, 2025).
Original GitHub issue: https://github.com/vxcontrol/pentagi/issues/30

The local deepseek model uses X-Ais-Token authentication. How should I set the model's authentication method in the pentagi configuration?

Originally created by @JacksonLvX on GitHub (Apr 8, 2025). Original GitHub issue: https://github.com/vxcontrol/pentagi/issues/30 The local deepseek model uses X-Ais-Token authentication. How should I set the model's authentication method in the pentagi configuration?

yindo closed this issue

2026-06-06 22:07:58 -04:00

yindo commented

2026-06-06 22:07:59 -04:00

@asdek commented on GitHub (Apr 9, 2025):

Hello @JacksonLvX,

Thank you for reaching out.

I'm not familiar with the X-Ais-Token authentication you mentioned. Could you please provide more details about the LLM backend you're using so I can better assist you?

@asdek commented on GitHub (Apr 9, 2025): Hello @JacksonLvX, Thank you for reaching out. I'm not familiar with the X-Ais-Token authentication you mentioned. Could you please provide more details about the LLM backend you're using so I can better assist you?

yindo commented

2026-06-06 22:07:59 -04:00

@icebeer0525 commented on GitHub (Apr 10, 2025):

In DeepSeek's local deployment, the X-Ais-Token is a custom header used for authentication, typically passed via HTTP headers when calling the API. How to config it in pentagi?

@icebeer0525 commented on GitHub (Apr 10, 2025): In DeepSeek's local deployment, the X-Ais-Token is a custom header used for authentication, typically passed via HTTP headers when calling the API. How to config it in pentagi?

yindo commented

2026-06-06 22:07:59 -04:00

@JacksonLvX commented on GitHub (Apr 10, 2025):

Hello @asdek ,

url =https://127.0.0.1:1111/v1
headers = {
    'Content-Type':'application/json',
    'X-Ais-Token':'ACCESSCODE XXXXXX'
}
payload = json.dumps({
    "model":"deepseek-r1",
    "messages":[
        {
            "role":"user",
            "content":"hello"
        }
    ]
})

response = requests.requests("POST",url=url,headers=headers,data=payload)

I use this method to request a local deepseek model. How should I configure the pentagi tool so that it passes X-Ais-Token authentication when calling deepseek?

@JacksonLvX commented on GitHub (Apr 10, 2025): Hello @asdek , ```markdown url =https://127.0.0.1:1111/v1 headers = { 'Content-Type':'application/json', 'X-Ais-Token':'ACCESSCODE XXXXXX' } payload = json.dumps({ "model":"deepseek-r1", "messages":[ { "role":"user", "content":"hello" } ] }) response = requests.requests("POST",url=url,headers=headers,data=payload) ``` I use this method to request a local deepseek model. How should I configure the pentagi tool so that it passes X-Ais-Token authentication when calling deepseek?

yindo commented

2026-06-06 22:07:59 -04:00

@asdek commented on GitHub (Apr 23, 2025):

Hello @JacksonLvX and @icebeer0525

I apologize for the delayed response; I missed the notification about your messages.

Currently, PentAGI works only with OpenAI-compatible LLM servers. If you're considering an on-premises backend, you might want to explore vLLM, which is tested and supported within our framework.

According to the OpenAI API specification, authorization is handled via the header:

Authorization: Bearer TOKEN

At this time, this is the only method PentAGI supports for sending authentication data.

I searched online for any backend that supports the X-Ais-Token header but unfortunately couldn't find any information. Could you please share a link to the backend you're using and its documentation? This would help me better understand how to integrate it with PentAGI.

It's possible that your backend doesn't fully implement the OpenAI API specification. If that's the case, the issue might not only involve passing authentication data but also the overall communication with the LLM, including how messages are sent and received.

Without compatibility with the OpenAI API standards, PentAGI may not be able to interact properly with your LLM server.

Please provide more details about your backend so I can assist you further.

@asdek commented on GitHub (Apr 23, 2025): Hello @JacksonLvX and @icebeer0525 I apologize for the delayed response; I missed the notification about your messages. Currently, PentAGI works only with **OpenAI-compatible LLM servers**. If you're considering an on-premises backend, you might want to explore [vLLM](https://github.com/vllm-project/vllm), which is tested and supported within our framework. According to the OpenAI API specification, authorization is handled via the header: ``` Authorization: Bearer TOKEN ``` At this time, this is the only method PentAGI supports for sending authentication data. I searched online for any backend that supports the `X-Ais-Token` header but unfortunately couldn't find any information. Could you please share a link to the backend you're using and its documentation? This would help me better understand how to integrate it with PentAGI. It's possible that your backend doesn't fully implement the OpenAI API specification. If that's the case, the issue might not only involve passing authentication data but also the overall communication with the LLM, including how messages are sent and received. Without compatibility with the OpenAI API standards, PentAGI may not be able to interact properly with your LLM server. Please provide more details about your backend so I can assist you further.

yindo referenced this issue

2026-06-06 22:09:16 -04:00

[PR #13] [MERGED] chore(deps): bump golang.org/x/net from 0.33.0 to 0.36.0 in /backend in the go_modules group across 1 directory #119

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: vxcontrol/pentagi#13