Use Kimi K2.6 with Continue in Morocco — velqa.dev, pay in dirhams (MAD)

Continue is a VS Code / JetBrains extension for code autocomplete and chat. Configure it to use velqa.dev models with payment in Moroccan dirhams.

Prerequisites

• VS Code with the Continue extension installed
• A velqa.dev API key (see the quickstart)

Configuration (config.yaml — current format)

Continue uses ~/.continue/config.yaml as its main configuration file. Add a block to the models section:

models:
  - name: Kimi K2.6 (Velqa)
    provider: openai
    model: kimi-k2.6
    apiBase: https://api.velqa.dev/v1
    apiKey: sk-...
    defaultCompletionOptions:
      contextLength: 131072
      maxTokens: 16384
    roles:
      - chat
      - edit

  - name: MiniMax M3 (Velqa)
    provider: openai
    model: minimax-m3
    apiBase: https://api.velqa.dev/v1
    apiKey: sk-...
    defaultCompletionOptions:
      contextLength: 131072
      maxTokens: 16384
    roles:
      - chat
      - edit
      - autocomplete

defaultCompletionOptions.contextLength and maxTokens are important: they cap the context size Continue sends and trigger truncation/compaction before overflow. Without them, a long session lets the context grow and eventually triggers Trop de tokens demandés. Reessaie plus tard., while burning through your quota faster. Use the model's real window: kimi-k2.6/minimax-m3/glm-4.7 131072, hy3 262144, glm-4.7-flash 200000, deepseek-v4-flash 65536 (see velqa.dev/models).

Configuration (config.json — deprecated format)

If you still use the old JSON format (~/.continue/config.json):

{
  "models": [
    {
      "name": "Kimi K2.6 (Velqa)",
      "provider": "openai",
      "model": "kimi-k2.6",
      "apiBase": "https://api.velqa.dev/v1",
      "apiKey": "sk-..."
    },
    {
      "name": "MiniMax M3 (Velqa)",
      "provider": "openai",
      "model": "minimax-m3",
      "apiBase": "https://api.velqa.dev/v1",
      "apiKey": "sk-..."
    }
  ]
}

The YAML format is recommended for new installations.

Test

After saving the file, open the Continue panel in VS Code. Select Kimi K2.6 (Velqa) in the model selector and send a message. A valid response confirms the connection.

Models available on velqa.dev

See velqa.dev/models for the full list and prices in MAD.