Skip to content

Cline: using the VIDAI Control Plane as the backend

Cline's API Provider settings include an "OpenAI Compatible" option (and a native Anthropic option) that accept a custom base URL and API key. Point either at the control plane and every autonomous coding step Cline runs flows through us.

TL;DR — OpenAI Compatible

  1. Cline sidebar → Settings icon → API Provider.
  2. Choose OpenAI Compatible.
  3. Fill in:
  4. Base URL: https://your-vidai-server.example.com/v1
  5. API Key: your VIDAI API key
  6. Model ID: a model registered on the control plane (e.g. gpt-4o-mini)
  7. Done. Start a task.

TL;DR — Anthropic (Claude Agent SDK path)

  1. API Provider → Anthropic.
  2. Fill in:
  3. Anthropic Base URL: https://your-vidai-server.example.com (no /v1 — Cline appends it)
  4. API Key: your VIDAI API key
  5. Model ID: claude-haiku-4-5 (or any Claude-family model registered on the control plane)

Prerequisites

  • Cline installed as a VS Code extension.
  • Control plane base URL and an API key from API Keys.
  • A model registered on the Models page.

What runs through the control plane

Cline is autonomous — every reasoning step, file read, edit attempt, tool invocation, and approval prompt makes an LLM call. Every one of those is a control-plane request that shows up on Request Logs.

Attribution

One VIDAI API key per developer keeps Cline spend attributed per seat. If Cline traffic dominates your Chargeback totals (it often does — autonomous coding is chatty), a dedicated application on the Applications page for "Cline usage" gives you a clean line item.

Verify it works

Start a new task with the message:

reply with exactly: ok

Cline will make at least one control-plane request before attempting anything else. A row appears on Request Logs.

If something's off

Raise an issue at github.com/vidaiUK/vidai-quickstart/issues with your API Provider config (masking the key) and the Cline version. We'll get it sorted.

Where to go next