Integration
Windsurf with a custom OpenAI-compatible provider
Windsurf supports custom model providers in recent versions. The setup is the same as any OpenAI-compatible endpoint: base URL, key, model ID.
Updated
Setup
Open Windsurf's model/provider settings and add an OpenAI-compatible provider. Use https://llmcode.io/v1 as the base URL and the published slug as the model.
Menu names change between releases. If your version has no custom provider field, update Windsurf or use an extension that supports custom base URLs.
Keys and limits
Use a separate key for the editor with a spend limit and expiry. The key is stored on the machine.
For autocomplete, pick a fast profile; for chat and refactors, pick a stronger one.
If it does not work
Test the key outside the editor: GET /v1/models. If that fails, the problem is the key or the network, not Windsurf.
Check the model ID: a display name such as “DeepSeek Flash” is not the same as the slug used in requests.
Connect
Provider type: OpenAI Compatible
Base URL: https://llmcode.io/v1
API key: sk-llmcode-…
Model: flashPublished model prices
USD per 1M tokens. The numbers come from the public catalog on every page rebuild — this is the sale price, not an estimate.
| Model | Input | Output | Cache |
|---|---|---|---|
| Claude Opus 5.5 | $2.70 | $8.00 | $0.14 |
| DeepSeek Flash | $0.50 | $2.00 | $0.01 |
| Claude Sonnet 5.5 | $1.33 | $6.67 | $0.07 |
| DeepSeek V4 Pro | $2.20 | $6.60 | $0.07 |
Top-up from $10, balance in USD. 1 USD = 1,000,000 UT. No subscription and no auto-charge.
Questions
Does Windsurf accept any OpenAI-compatible API?
Recent versions support custom providers, but the exact fields depend on the release. Check your version's settings.
Which model should I use for autocomplete?
The fastest published profile, with a separate key so autocomplete cannot exhaust the chat budget.
Is the key visible to others?
It is stored locally by the editor. Treat it as a secret and revoke it in the cabinet if the machine is lost.