MegaBrain University · guide

How to Use Claude Code Without a Subscription: Pay-per-Token Claude via One API Key

Updated November 11, 2026 · 5 min setup

What is Claude Code

Claude Code is Anthropic’s AI coding agent. It runs in the terminal, in VS Code and JetBrains, and headless in scripts and CI: it reads the codebase, edits files, runs commands and tests, uses subagents, hooks, skills and MCP servers.

Claude Code normally signs in with a Claude subscription or an Anthropic Console key, but it also supports an LLM gateway: any server that speaks the Anthropic Messages API, set with ANTHROPIC_BASE_URL. MegaBrain is one. You pay per token from one balance shared with every other model, and each developer gets a key whose usage you can see. Anthropic supports Claude Code only on Claude models, so this guide uses Claude Opus, Sonnet and Haiku.

Setup, step by step

1. Step 1. Get a MegaBrain API key

   Sign up with GitHub, Google or email, add funds on the Balance page and create a key on the API key page. The key works for every text, image and video model, and each request is paid from your balance — no subscription.

   Put the key in an environment variable so it never lands in a config file you share:

   # ~/.zshrc or ~/.bashrc
   export MEGABRAIN_API_KEY="your-megabrain-key"

   Copy

   curl https://getmegabrain.com/api/gateway/v1/chat/completions \
     -H "Authorization: Bearer $MEGABRAIN_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{"model": "anthropic/claude-sonnet-5.5", "messages": [{"role": "user", "content": "Say OK"}]}'

   Copy

   Check the key before you touch the agent. A JSON reply with choices means the key, the balance and the model all work:

2. Step 2. Install Claude Code

   On macOS, Linux or WSL:

   curl -fsSL https://claude.ai/install.sh | bash

   Copy

   On Windows (PowerShell): irm https://claude.ai/install.ps1 | iex. With npm: npm install -g @anthropic-ai/claude-code.

3. Step 3. Point Claude Code at MegaBrain

   Add an env block to ~/.claude/settings.json (it applies to every project and to background agents). The base URL has no /v1: Claude Code adds /v1/messages itself.

   {
     "env": {
       "ANTHROPIC_BASE_URL": "https://getmegabrain.com/api/anthropic",
       "ANTHROPIC_AUTH_TOKEN": "your-megabrain-key",
       "ANTHROPIC_MODEL": "anthropic/claude-sonnet-5.5",
       "ANTHROPIC_DEFAULT_OPUS_MODEL": "anthropic/claude-opus-5.5",
       "ANTHROPIC_DEFAULT_SONNET_MODEL": "anthropic/claude-sonnet-5.5",
       "ANTHROPIC_DEFAULT_HAIKU_MODEL": "anthropic/claude-haiku-5.5"
     }
   }

   Copy

   Before you start Claude Code, check the endpoint and key in one request:

   curl -sS -X POST "https://getmegabrain.com/api/anthropic/v1/messages" \
     -H "Authorization: Bearer $MEGABRAIN_API_KEY" \
     -H "anthropic-version: 2023-06-01" \
     -H "content-type: application/json" \
     -d '{"model": "anthropic/claude-haiku-5.5", "max_tokens": 20, "messages": [{"role": "user", "content": "Say OK"}]}'

   Copy

   ANTHROPIC_AUTH_TOKEN sends the key as Authorization: Bearer and takes over from a saved claude.ai login at once. Never put the key in a project’s .claude/settings.json: that file is committed. The ANTHROPIC_DEFAULT_*_MODEL variables decide what the opus, sonnet and haiku choices in /model call; Haiku also runs background tasks such as session titles.

4. Step 4. Start Claude Code and check

   Run claude, then /status: the Anthropic base URL line must show https://getmegabrain.com/api/anthropic and the credential line must name ANTHROPIC_AUTH_TOKEN. Send a prompt; a normal answer means it works. If you see a warning about two credential sources, run /logout to drop the old claude.ai login.

5. Step 5. VS Code, CI and scripts

   In the VS Code extension, put the same variables in claudeCode.environmentVariables in your user settings (JSON). In GitHub Actions, set ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN in the workflow env and pass the same secret as the action’s anthropic_api_key input. In scripts, claude -p "task" runs one task headless with the same settings.

Which model to choose

For Claude Code use Claude models: anthropic/claude-sonnet-5.5 for everyday coding, anthropic/claude-opus-5.5 (or anthropic/claude-fable-5.2) for hard refactors and architecture, anthropic/claude-haiku-5.5 for background tasks. Anthropic does not support Claude Code on other labs’ models; for GPT, Gemini, Kimi or GLM use Codex, OpenCode or Kilo Code with the same key.

Every model with its price and context window is in the catalog. They all work with the same key.

What you can do

Business use cases

Review the diff of this pull request. List: 1) bugs or regressions with file and line, 2) changes without tests, 3) anything touching auth, payments or data deletion. Be specific and short; skip style nits.

Troubleshooting

FAQ