Use AnyAPI in VS Code

AnyAPI works with any VS Code AI plugin that supports custom OpenAI-compatible endpoints. Point the plugin at https://api.anyapi.ai/v1, add your API key, and you get access to 300+ models.

Prerequisites

Get your API key from the AnyAPI Dashboard
Pick a model from the model catalog

All plugins below use the same connection details:

Parameter	Value
Base URL	`https://api.anyapi.ai/v1`
API Key	Your AnyAPI key (`sk-...`)
Model ID	e.g. `openai/gpt-4o`, `anthropic/claude-sonnet-4`

AnyAPI uses provider/model format for model IDs. Bare names like gpt-4o without the vendor prefix won’t work. Browse all models at anyapi.ai/ai-models.

Option 1: Cline (Recommended)

Autonomous AI coding agent. Edits files, runs terminal commands, browses the web. No sign-up required — setup takes 30 seconds.

Setup

Install Cline from the Extensions marketplace (Ctrl+Shift+X / Cmd+Shift+X)
Open the Cline panel in the sidebar
Click the gear icon (Settings)
Set API Provider to OpenAI Compatible
Fill in:

Field	Value
Base URL	`https://api.anyapi.ai/v1`
API Key	Your AnyAPI key
Model ID	e.g. `anthropic/claude-sonnet-4`

Click Save — you’re ready to go

Usage

Type a task in the chat (e.g. “Create a REST API with Express”) and Cline will autonomously create files, install dependencies, and write code. It asks for approval before each action.

Use large-context models like anthropic/claude-sonnet-4 for complex tasks and cheaper models like openai/gpt-4o-mini for quick fixes.

Option 2: Continue.dev

Open-source AI assistant with chat, autocomplete, refactoring, and code explanation.

Setup

Install Continue from the Extensions marketplace
Sign in with a Continue.dev account when prompted
Open ~/.continue/config.yaml (Cmd+Shift+P → Continue: Open configuration file)
Replace the contents with:

name: AnyAPI Config
version: 1.0.0
schema: v1
models:
  - name: AnyAPI — Claude Sonnet
    provider: openai
    model: anthropic/claude-sonnet-4
    apiBase: https://api.anyapi.ai/v1
    apiKey: sk-your-anyapi-key
    roles:
      - chat
      - edit

  - name: AnyAPI — GPT-4o
    provider: openai
    model: openai/gpt-4o
    apiBase: https://api.anyapi.ai/v1
    apiKey: sk-your-anyapi-key
    roles:
      - chat
      - edit

  - name: AnyAPI Autocomplete
    provider: openai
    model: openai/gpt-4o-mini
    apiBase: https://api.anyapi.ai/v1
    apiKey: sk-your-anyapi-key
    roles:
      - autocomplete

Reload VS Code (Cmd+Shift+P → Developer: Reload Window)

Newer versions of Continue require signing in with a Continue.dev account on first launch. If the login hangs, reload VS Code and try again.

Usage

Chat — Open the Continue panel, select a model, and ask questions
Autocomplete — Start typing code, press Tab to accept suggestions
Refactoring — Select code, press Ctrl+Shift+L / Cmd+Shift+L and describe the change

Option 3: GitHub Copilot BYOK

Use AnyAPI models inside GitHub Copilot Chat with Bring Your Own Key.

Requires VS Code Insiders and a GitHub Copilot subscription (Free plan or above). If the Chat: Manage Language Models command is not found, download VS Code Insiders.

Setup

Open Command Palette (Cmd+Shift+P / Ctrl+Shift+P)
Type Chat: Manage Language Models
Click Add Models → Custom Endpoint
Configure the provider:

Field	Value
API Key	Your AnyAPI key
API Type	`Chat Completions`

Add a model:

Field	Value
ID	`anthropic/claude-sonnet-4`
Name	`AnyAPI Claude Sonnet`
URL	`https://api.anyapi.ai/v1/chat/completions`

Save — the model appears in the Copilot Chat model picker

This creates a chatLanguageModels.json file:

[
  {
    "vendor": "customendpoint",
    "apiKey": "sk-your-anyapi-key",
    "apiType": "chat-completions",
    "models": [
      {
        "id": "anthropic/claude-sonnet-4",
        "name": "AnyAPI Claude Sonnet",
        "url": "https://api.anyapi.ai/v1/chat/completions",
        "toolCalling": true,
        "maxInputTokens": 200000,
        "maxOutputTokens": 16000
      }
    ]
  }
]

Copilot BYOK only applies to chat and agent mode. Inline autocomplete still uses GitHub’s own models.

Option 4: VS Code Native Custom Endpoint

Built-in support for custom AI endpoints, no extensions needed.

Currently available in VS Code Insiders only.

Setup

Open VS Code Insiders
Click the model dropdown in Copilot Chat
Select Manage Language Models…
Click Add Models → Custom Endpoint
Fill in:

Field	Value
Group Name	`AnyAPI`
Display Name	`GPT-4o via AnyAPI`
Endpoint URL	`https://api.anyapi.ai/v1/chat/completions`
API Key	Your AnyAPI key
API Type	`Chat Completions`

Option 5: LiteLLM Extensions

Two extensions designed for LiteLLM proxies — fully compatible with AnyAPI.

LiteLLM Connector for Copilot

Install LiteLLM Connector for Copilot
Open Settings (Ctrl+, / Cmd+,) and search for litellm
Set:

Setting	Value
Base URL	`https://api.anyapi.ai`
API Key	Your AnyAPI key
Default Model	`openai/gpt-4o`

LiteLLM Provider for Copilot Chat

Install LiteLLM Provider for Copilot Chat
Cmd+Shift+P → LiteLLM: Add Server
Enter label AnyAPI, URL https://api.anyapi.ai, and your API key

Supports multiple servers, automatic model aggregation, and cheapest/fastest model selection.

Option 6: vscode-openai

Lightweight extension for OpenAI-compatible APIs.

Install vscode-openai
Cmd+Shift+P → vscode-openai: Set API URL → https://api.anyapi.ai/v1
Cmd+Shift+P → vscode-openai: Set API Key → your AnyAPI key

Comparison

Plugin	Chat	Autocomplete	Agent	Free	Requires
Cline	Yes	No	Yes	Yes	—
Continue.dev	Yes	Yes	No	Yes	Continue account
Copilot BYOK	Yes	No	Yes	No	Copilot + VS Code Insiders
VS Code Native	Yes	No	Yes	Yes	VS Code Insiders
LiteLLM Extensions	Yes	No	No	No	Copilot
vscode-openai	Yes	No	No	Yes	—

Recommendations

Quick start, no sign-up — Cline
Autocomplete + chat — Continue.dev
Already use Copilot — BYOK or LiteLLM extensions
No extensions — VS Code Native Custom Endpoint (Insiders only)

Recommended Models for Coding

Model	Best for
`anthropic/claude-sonnet-4`	Fast edits, everyday coding
`anthropic/claude-opus-4`	Complex architecture, deep reasoning
`openai/gpt-4o`	Multi-modal, broad knowledge
`openai/gpt-4o-mini`	Autocomplete, quick tasks
`deepseek/deepseek-r1`	Algorithms, math

Troubleshooting

”Model not found”

Use the full provider/model format. Check available models:

curl https://api.anyapi.ai/v1/models \
  -H "Authorization: Bearer sk-your-anyapi-key"

“Unauthorized” (401)

Verify your API key is correct and active in the AnyAPI Dashboard.

”Payment Required” (402)

Top up your balance or check the limits on your key.

Slow responses

Use a faster model (openai/gpt-4o-mini)
Increase the timeout in plugin settings
Reduce maxTokens

Next Steps

Cline Guide

Quick setup guide for Cline

Continue.dev Guide

Quick setup guide for Continue.dev

All Integrations

Browse all supported tools

Model Catalog

Browse 300+ available models

Get started

Features

Use Cases

Developer guides

Integrations

API Reference

VS Code Integration

Use AnyAPI in VS Code

Prerequisites

Option 1: Cline (Recommended)

Setup

Usage

Option 2: Continue.dev

Setup

Usage

Option 3: GitHub Copilot BYOK

Setup

Option 4: VS Code Native Custom Endpoint

Setup

Option 5: LiteLLM Extensions

LiteLLM Connector for Copilot

LiteLLM Provider for Copilot Chat

Option 6: vscode-openai

Comparison

Recommendations

Recommended Models for Coding

Troubleshooting

”Model not found”

“Unauthorized” (401)

”Payment Required” (402)

Slow responses

Next Steps

Cline Guide

Continue.dev Guide

All Integrations

Model Catalog

​Use AnyAPI in VS Code

​Prerequisites

​Option 1: Cline (Recommended)

​Setup

​Usage

​Option 2: Continue.dev

​Setup

​Usage

​Option 3: GitHub Copilot BYOK

​Setup

​Option 4: VS Code Native Custom Endpoint

​Setup

​Option 5: LiteLLM Extensions

​LiteLLM Connector for Copilot

​LiteLLM Provider for Copilot Chat

​Option 6: vscode-openai

​Comparison

​Recommendations

​Recommended Models for Coding

​Troubleshooting

​”Model not found”

​“Unauthorized” (401)

​”Payment Required” (402)

​Slow responses

​Next Steps

Cline Guide

Continue.dev Guide

All Integrations

Model Catalog

Use AnyAPI in VS Code

Prerequisites

Option 1: Cline (Recommended)

Setup

Usage

Option 2: Continue.dev

Setup

Usage

Option 3: GitHub Copilot BYOK

Setup

Option 4: VS Code Native Custom Endpoint

Setup

Option 5: LiteLLM Extensions

LiteLLM Connector for Copilot

LiteLLM Provider for Copilot Chat

Option 6: vscode-openai

Comparison

Recommendations

Recommended Models for Coding

Troubleshooting

”Model not found”

“Unauthorized” (401)

”Payment Required” (402)

Slow responses

Next Steps