Use sintaks.dev AI gateway models inside VS Code chat — 300+ OpenAI-compatible models (OpenAI, Anthropic, Google, Alibaba, and more) billed per request in IDR with QRIS top-ups. No credit card required.
Features
Registers a sintaks model vendor in the VS Code model picker (Agents view / Copilot Chat)
Discovers models live from GET https://api.sintaks.dev/v1/models (chat models only, with context window and capability metadata)
Streaming chat completions via POST /v1/chat/completions
Function/tool calling, vision input (for supports_vision models), and reasoning_effort passthrough
API key stored in VS Code Secret Storage — never written to settings files
Install the extension (.vsix or from the Marketplace).
Open the Command Palette (Cmd/Ctrl+Shift+P) and run Chat: Manage Language Models.
(Or, from the chat model picker in the Agents view, click the gear icon → Manage Language Models.)
In the Language Models editor, enable sintaks.dev models.
When prompted, paste an API key from app.sintaks.dev.
You can also run sintaks.dev: Manage Provider from the Command Palette to
set/clear the key, refresh the model list, or open the dashboard.
Pick any sintaks model in the model picker and chat.
Configuration
Setting
Default
Description
sintaks.baseUrl
https://api.sintaks.dev/v1
Gateway base URL (OpenAI-compatible /v1 root). Point at a local gateway for development.
Development
pnpm install
pnpm run compile # bundle to dist/extension.js
pnpm run typecheck # strict tsc check
pnpm run watch # rebuild on change
Press F5 in VS Code to launch an Extension Development Host with the provider loaded.
Package a .vsix for sideloading:
pnpm run vsce
Requirements
VS Code ^1.104.0 (stable languageModelChatProviders contribution point)
GitHub Copilot Chat (built into recent VS Code releases; install it from the Marketplace on older versions) and a GitHub Copilot account on an individual plan — models from third-party providers are only surfaced to individual Copilot plans
A sintaks.dev account with microcredit balance for non-free models