Continue (continue.dev): using the VIDAI Control Plane as the backend¶

Continue's models list in config.json supports an OpenAI-compatible provider. Add an entry pointing at the control plane and every chat, autocomplete, edit, and agent run in your IDE flows through us.

TL;DR¶

Edit ~/.continue/config.json (or the workspace equivalent):

{
  "models": [
    {
      "title": "VIDAI Control Plane · GPT-4o mini",
      "provider": "openai",
      "model": "gpt-4o-mini",
      "apiBase": "https://your-vidai-server.example.com/v1",
      "apiKey": "your-vidai-key"
    }
  ],
  "tabAutocompleteModel": {
    "title": "VIDAI Control Plane · autocomplete",
    "provider": "openai",
    "model": "gpt-4o-mini",
    "apiBase": "https://your-vidai-server.example.com/v1",
    "apiKey": "your-vidai-key"
  }
}

Reload Continue in your IDE (VS Code / JetBrains). Your VIDAI model appears in the model picker.

Prerequisites¶

Continue installed (VS Code extension or JetBrains plugin).
Control plane base URL and an API key from API Keys.
A model registered on the Models page.

Multiple models¶

Add several entries — Continue lists all of them in its model picker:

"models": [
  { "title": "GPT-4o mini (VIDAI)",   "provider": "openai", "model": "gpt-4o-mini",       "apiBase": "...", "apiKey": "..." },
  { "title": "Claude Sonnet (VIDAI)", "provider": "openai", "model": "claude-sonnet-4.5", "apiBase": "...", "apiKey": "..." },
  { "title": "Gemini Flash (VIDAI)",  "provider": "openai", "model": "gemini-2.0-flash",  "apiBase": "...", "apiKey": "..." }
]

Autocomplete separately¶

tabAutocompleteModel is separate from the chat/edit list. Point it at a cheap model — autocomplete is high-volume and low per-request:

"tabAutocompleteModel": {
  "provider": "openai",
  "model": "gpt-4o-mini",
  "apiBase": "https://your-vidai-server.example.com/v1",
  "apiKey": "your-vidai-key"
}

Attribution¶

One VIDAI API key per developer is the clean pattern. Autocomplete alone can generate hundreds of requests per hour per active seat — putting autocomplete on its own key (or a dedicated application on the Applications page) keeps it visible on Chargeback as its own line item.

Verify it works¶

In Continue, open the chat panel, select your VIDAI model, and send:

reply with exactly: ok

A row appears on Request Logs.

If something's off¶

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