Use Continue.dev with a cheap OpenAI-compatible API
Continue.dev is a popular open-source AI assistant for VS Code and JetBrains. Like most modern AI coding tools, it speaks the OpenAI-compatible API format — so you don't have to pay official list price. Point it at a cheaper compatible provider and everything works unchanged.
This guide shows the exact Continue config to use an OpenAI-compatible gateway like APIVAI.
Why it works
Continue lets you define any model with a provider: "openai" block and a custom apiBase. That's all it takes to route through a discount gateway — the requests are standard OpenAI chat completions.
Configure Continue
Open your Continue config (~/.continue/config.json, or the YAML config in newer versions) and add a model:
{
"models": [
{
"title": "APIVAI - Claude Sonnet",
"provider": "openai",
"model": "claude-sonnet-4-6",
"apiBase": "https://api.apivai.com/v1",
"apiKey": "sk-your-apivai-key"
}
]
}
That's it. Reload Continue and the model appears in the dropdown. Send a message to confirm it streams.
Pick a model that exists
Model availability varies, so list what's available first instead of guessing:
curl -s https://api.apivai.com/v1/models \
-H "Authorization: Bearer $APIVAI_API_KEY"
Use a model name from that response for the model field.
Things to verify
- Streaming — Continue streams responses; make sure your provider passes real SSE chunks so output appears token by token.
-
Model name — must match one the provider serves (avoids
model_not_found). - Context window — pick a model whose context fits your codebase chunks.
Why APIVAI
APIVAI is OpenAI- and Anthropic-compatible, covers Claude and GPT models behind one key, and is priced at a fraction of official list. Pay-as-you-go with crypto, USDT, or Alipay, no VPN required — handy if you can't or don't want to use a card.
Get started
Grab a key, drop the config block above into Continue with a model from /v1/models, and start coding at a fraction of the cost — no code changes, same workflow.
Top comments (0)