Light
Client & Key Setup

Connect Apps and API Keys to Light API

Use a personal API key with a spending cap in your code editor, terminal or custom application. Choose between manual endpoint configuration and ready installation archives for Windows and macOS.

Open Connection Wizard Check Available Models HTTP Request Format

Step-by-Step Setup

  1. Sign in to your account: authenticate via Telegram (shared wallet with the Telegram bot) or with a website username and password (independent website wallet).
  2. Activate your wallet: creating API keys requires an activated wallet after a first balance credit. Public crypto top-ups on the website are currently disabled (pre-launch mode); contact support for a test credit to activate your wallet.
  3. Create a capped key: choose a key name, spending limit in USD (from $0.01), expiration period (7, 30, 90, 365 days or no expiration) and allowed models. A key cap limits spending from your wallet balance and does not add funds.
  4. Select your setup method: copy the secret key and gateway endpoint https://api.lightapi.app/v1 for manual integration, or download a preconfigured ZIP installer for your operating system and application.
  5. Run the installer: extract the entire archive into a folder and run Install.cmd on Windows or Install.command on macOS.

Supported Applications

Installation bundles are tailored to your selected API key, operating system and default model.

VS Code (Windows & macOS)

Configures the Continue extension in an isolated profile and remembers your selected project folder. On Windows, the launcher shows installation progress and connection verification results.

Cursor via Continue

The installer opens a configured Cursor profile with the Continue extension. Pick models inside Continue (sidebar icon or F1 -> Continue: Focus Continue Chat). The built-in native Cursor chat is separate, requires public HTTPS and must be configured manually.

Claude Code (Terminal CLI)

Command-line engineering assistant in your terminal, configured to use the Anthropic-compatible /v1/messages endpoint with your personal key and chosen model.

OpenCode (Terminal CLI)

Terminal coding tool connected over the OpenAI-compatible /v1/chat/completions endpoint with support for reading and editing project files.

Models Available for API Keys & Installers

The following models are currently active for API connection in the server configuration:

  • GPT 5.6 Sol (gpt-5.6-sol, OpenAI)
  • GPT 6 Astra (gpt-6-astra, OpenAI)
  • Claude Sonnet 5.5 (claude-sonnet-5-5, Anthropic)
  • Opus 5.5 - supplier 1 (claude-opus-5-5-provider-1, Anthropic)
  • GLM 5.3 (glm-5.3, Z.ai)
  • Kimi K3 (kimi-k3, Moonshot AI)
  • DeepSeek V4 Pro (deepseek-v4-pro, DeepSeek)
  • DeepSeek V4.1 Flash (deepseek-v4.1-flash, DeepSeek)
  • Qwen 3.8 Flash (qwen3.8-flash, Qwen)
  • MiMo V2.6 Pro (mimo-v2.6-pro, MiMo)
  • Grok 4.7 (grok-4.7, xAI)
  • Gemini 3.8 Flash (gemini-3.8-flash, Google)
  • HY4 Preview (hy4-preview, HY)
  • Gemini 4 Argon (gemini-4-argon, Google)

In unified token billing mode, flat-enabled models use the -tokens suffix according to your wallet mode.

Key Lifecycle & Security

Replacing Caps & Allowed Models

Updating a key's spending cap, expiration or allowed models creates a new replacement key and automatically revokes the previous one. Your wallet balance and usage history remain intact; simply update the key in your app or download a fresh installer archive.

Revocation & Secret Protection

Each account can hold up to 10 active keys at a time. You can request revocation from your account console. New requests are blocked once the gateway applies it; a request already in progress may finish. Downloaded installer archives contain your private API key - keep them confidential.