Setup LLM Provider in VS Code Extensions (Cline, Kiocode)

All models are accessed through Cloudflare AI Gateway (demo-hkmci). Users only need the gateway API token — no direct provider keys required.

👤 User Guide: Quick Model Reference

Gateway API Token (same for all remote models)

API Token: [Ask your Cloudflare admin]

⚠️ Local vs Remote Base URL

Type Base URL pattern Example
Remote (Cloudflare Gateway) https://gateway.ai.cloudflare.com/v1/{account}/{gateway}/{provider-path} See table below
Local (Ollama, LM Studio) http://localhost:{port}/v1 http://localhost:11434/v1

Use the remote URLs below when connected to the internet via the Cloudflare gateway. Switch to a local URL if running models on your own machine.

Available Models

Cloudflare Workers AI

Provider Type: OpenAI Compatible Base URL:

https://gateway.ai.cloudflare.com/v1/b326904912840c25f63808a1d1e479aa/demo-hkmci/workers-ai/v1
Model Model ID
Nvidia Nemotron 3 120B @cf/nvidia/nemotron-3-120b-a12b
Moonshot Kimi K2.5 @cf/moonshotai/kimi-k2.5
GLM 4.7 Flash @cf/zai-org/glm-4.7-flash

Google Gemini (via Google AI Studio)

Provider Type: OpenAI Compatible Base URL:

https://gateway.ai.cloudflare.com/v1/b326904912840c25f63808a1d1e479aa/demo-hkmci/google-ai-studio/v1beta/openai
Model Model ID Notes
Gemini 3.1 Pro gemini-3.1-pro New, Preview
Gemini 3 Flash gemini-3.0-flash Preview
Gemini 3.1 Flash-Lite gemini-3.1-flash-lite New, Preview

Preview model IDs may change — verify at aistudio.google.com

Anthropic Claude (Under Investigation)

Two paths are available — choose based on what the admin has configured:

Option A — Direct Anthropic via Gateway (requires Anthropic BYOK key in gateway) Provider Type: Anthropic Base URL:

https://gateway.ai.cloudflare.com/v1/b326904912840c25f63808a1d1e479aa/demo-hkmci/anthropic

Option B — Claude via GCP Vertex AI (requires GCP billing enabled + Claude access approved) ⚠️ Pending Provider Type: Anthropic Base URL:

https://gateway.ai.cloudflare.com/v1/b326904912840c25f63808a1d1e479aa/demo-hkmci/google-vertex-ai/v1/projects/mcps-testing-cloudflare-ai/locations/us-east5/publishers/anthropic/models
Model Model ID
Claude Opus 4.6 claude-opus-4-6
Claude Sonnet 4.6 claude-sonnet-4-6

Note: Claude models on GCP Vertex AI are Anthropic’s models hosted by Google — they are NOT Google’s own models. Vertex AI billing must be enabled and Claude access must be approved in Vertex AI Model Garden before these become available.

Screenshots

Kilocode: Kilocode Provider Setup

Cline: Cline Provider Setup

🔧 Admin Guide: Gateway Setup

Architecture Overview

Cline / Kiocode
      │
      │  cfut_ token (gateway auth)
      ▼
Cloudflare AI Gateway (demo-hkmci)
      │
      ├── /workers-ai/v1        → Cloudflare Workers AI (built-in)
      ├── /google-ai-studio/... → Google Gemini  [BYOK: Google AI Studio key]
      ├── /anthropic            → Anthropic Claude [BYOK: Anthropic key]
      └── /google-vertex-ai/... → GCP Vertex AI  [BYOK: Service account JSON] ⚠️ billing required

No provider keys are exposed to users — all credentials are stored as Provider Keys (BYOK) in the gateway.

Gateway Details

Setting Value
Account Master Concept Demo (mcmsp.dev)
Account ID b326904912840c25f63808a1d1e479aa
Gateway Name demo-hkmci
Authentication Enabled
Gateway Token cfut_... (distribute to users)

Provider Keys (BYOK) Configuration

Navigate to: Cloudflare Dashboard → AI → AI Gateway → demo-hkmci → Provider Keys

Provider Alias Key Source Status
Google AI Studio default aistudio.google.com/apikey ✅ Configured
Google Vertex AI default GCP service account JSON ⚠️ Billing not enabled
Anthropic default console.anthropic.com ⬜ Not configured

Always set alias to default so the gateway injects it automatically without requiring users to hold provider keys.

How to Deliver Config to Users

Send users the following (nothing else — no provider keys):

Cloudflare AI Gateway – VS Code Setup

Gateway API Token: cfut_...

For model-specific Base URLs and setup instructions, see:
https://sharehub.zorro.hk/documents/2026-04-01-vscode-llm-provider-setup.html

GCP Vertex AI Setup

The Vertex AI service account has been created and the key is stored in the gateway. Blocked on billing.

  • GCP Project: mcps-testing-cloudflare-ai
  • Service Account: cloudflare-ai-gateway-vertex@mcps-testing-cloudflare-ai.iam.gserviceaccount.com
  • Action required: Enable billing at https://console.developers.google.com/billing/enable?project=mcps-testing-cloudflare-ai

Once billing is enabled, Vertex AI models (Gemini Enterprise tier) will be available via:

https://gateway.ai.cloudflare.com/v1/b326904912840c25f63808a1d1e479aa/demo-hkmci/google-vertex-ai/v1/projects/mcps-testing-cloudflare-ai/locations/us-central1/publishers/google/models

Adding a New Provider

  1. Get the API key for the provider
  2. Go to demo-hkmci → Provider Keys → find provider → click +
  3. Paste the key, set alias to default, save
  4. Update the User Guide table above with the new Base URL and model IDs
  5. Republish this note: /kf-cli:publish 2026-04-01-vscode-llm-provider-setup.md

Captured: 2026-04-01 Last Updated: 2026-04-01