[Howto]: "How to Configure GPT Pilot to Use a Zscaler Certificate to Avoid SSL Errors on a Corporate Network?

[nisan221292]

Version

VisualStudio Code extension

Operating System

Windows 10

Your question

I am using GPT Pilot on my company laptop within a restricted corporate network. Due to multiple firewall restrictions, I am unable to establish a connection to LLM provider. To bypass this, I need to use a Zscaler certificate, which allows the connection to the LLM (OpenAI) to be accepted.

I would like to know how to configure GPT Pilot Code base to use this specific certificate for every session, so that the SSL error no longer appears.

[kalalxy]

There is an 'extra' param on llm config, you may load your certificate file like
"extra": { "ssl_cert_file": "path/to/your/zscaler-certificate.pem" }
besides,you need to modify the core/llm/openai_client.py code to support this config:
`
def _init_client(self):
# check ssl certification
ssl_cert_file = None
if self.config.extra and 'ssl_cert_file' in self.config.extra:
ssl_cert_file = self.config.extra['ssl_cert_file']

    # create client
    if ssl_cert_file:
        http_client = httpx.AsyncClient(
            verify=ssl_cert_file,
            timeout=Timeout(
                max(self.config.connect_timeout, self.config.read_timeout),
                connect=self.config.connect_timeout,
                read=self.config.read_timeout,
            ),
        )
        self.client = AsyncOpenAI(
            api_key=self.config.api_key,
            base_url=self.config.base_url,
            http_client=http_client,
        )
    else:
        self.client = AsyncOpenAI(
            api_key=self.config.api_key,
            base_url=self.config.base_url,
            timeout=Timeout(
                max(self.config.connect_timeout, self.config.read_timeout),
                connect=self.config.connect_timeout,
                read=self.config.read_timeout,
            ),
        )

# left unchanged

`