Cline
This guide explains how to configure Cline to work with the PwC AI-CoE Coding Agents Gateway.
Overview
Cline is a VS Code extension that supports multiple LLM providers. This guide covers configuration using the OpenAI-compatible API endpoint.
Base URL Configuration
In Cline's API Configuration panel, set the Base URL to:
https://idi-coding-agents.pwc.it
Cline automatically appends /chat/completions to this URL when making requests.
The endpoint https://idi-coding-agents.pwc.it is only accessible via PwC VPN set to Italy or Italy On-Premise.
An alternative endpoint (https://doc-interact-backend.pwc.it) is available on the public internet and can also be used as the Base URL, but is not recommended as it's protected by the Imperva firewall, which may block frequent requests from coding agents, interpreting them as DDoS attempts.
API Key Configuration
In the OpenAI Compatible API Key field, enter your credentials in the following format:
apikey=<your-api-key>&tenantid=<your-tenant-id>
Replace <your-api-key> and <your-tenant-id> with the values obtained as described in Get Access.
The gateway parses this string from the Authorization: Bearer header that Cline sends automatically.
Some Cline versions have a known issue where Custom Headers are intermittently dropped from requests, causing authentication failures. Using the API Key field with the apikey=...&tenantid=... format is more reliable because it goes through the standard Authorization header, which is always sent.
Custom Headers
In the Custom Headers section, add the following entry:
| Header | Value |
|---|---|
model-name | Model identifier — accepts any of the three formats: vendor naming convention (e.g., claude-sonnet-4-6), Coding Agents Gateway naming convention (e.g., GENAI_SHARED_VERTEXAI_ANTHROPIC_CLAUDE_46_SONNET), or PwC GenAI Shared Service naming convention (e.g., vertex_ai.anthropic.claude-sonnet-4-6). See Available Models for all values. |
You no longer need api-key, tenant-id, or auth-type custom headers if you configured the API Key field as described above.
Model Configuration
Cline does not auto-detect model limits. You must manually set Max Output Tokens (and Context Window Size) in the MODEL CONFIGURATION section, otherwise Cline will use incorrect defaults and may truncate responses or fail.
In the MODEL CONFIGURATION section, set the following parameters to match your chosen model:
| Parameter | Value |
|---|---|
| Context Window Size | Value from the Available Models table |
| Max Output Tokens | Value from the Available Models table |
Optional: Cost Tracking
To monitor conversation costs, you can also configure:
| Parameter | Value |
|---|---|
| Input Price / 1M Tokens | Value from the Available Models table |
| Output Price / 1M Tokens | Value from the Available Models table |
Leave the Temperature parameter at its default value for optimal results.