Table of contents

• Why one router
  • Claude's token efficiencies
• Getting an OpenRouter key
• Claude Code
• OpenCode
  • What didn't work
  • What worked
• OpenClaw
• Custom OpenRouter Usage MCP
  • Registering it
  • Usage output
  • Model pricing snapshot

Why one router

I run three AI tools regularly: Claude Code, OpenCode, and OpenClaw. Each has its own API key, its own billing dashboard, and its own model config. Anthropic bills Claude Code. Google bills OpenClaw's Gemini usage. OpenCode can talk to either.

That's three places to check spend, three sets of rate limits, and no single view of what I'm burning through. OpenRouter fixes this. It's a proxy that sits between your tools and the model providers. You get one API key, one billing dashboard, and the ability to swap models without changing provider credentials.

The tradeoff is a small markup on token costs and an extra network hop. For a personal setup where I want visibility and flexibility, that's fine.

Claude's token efficiencies

Anthropic's Pro and Max plans include Claude Code usage at a steep discount compared to raw API pricing. Pro is $20/month with generous token allowances. Max plans scale up from there. If you only use Claude Code, the Pro plan is the cheapest way to run it.

The problem is that those plan tokens only work with Claude Code through Anthropic's direct API. OpenCode and OpenClaw are third-party tools. They can call Claude models, but not through your Pro subscription. They hit the standard API and bill at full rate.

So the choice is: keep Claude Code on the Pro plan (cheapest for that one tool) and manage the other two separately, or route everything through OpenRouter and get unified billing at API rates. I went with OpenRouter because the visibility across all three tools matters more to me than the per-token savings on Claude Code alone.

Getting an OpenRouter key

1. Sign up at openrouter.ai
2. Go to Settings > Keys and create a new API key
3. Add credits (pay-as-you-go, no minimum)

OpenRouter gave me $100 in free credits as a signup promo. That's a lot of tokens to experiment with before committing any money.

Set the key in your shell:

export OPENROUTER_API_KEY="sk-or-v1-..."

Add it to your ~/.zshrc or ~/.bashrc so it persists. The same key works for every tool below.

OpenRouter's models page shows per-token pricing for every model. Useful for comparing costs before swapping models around.

Claude Code

Claude Code normally talks directly to Anthropic's API. To route it through OpenRouter, you override the base URL and auth token via environment variables:

ANTHROPIC_BASE_URL="https://openrouter.ai/api" \
ANTHROPIC_AUTH_TOKEN="$OPENROUTER_API_KEY" \
ANTHROPIC_API_KEY="" \
CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 \
claude

Four variables. ANTHROPIC_BASE_URL points Claude Code at OpenRouter instead of Anthropic. ANTHROPIC_AUTH_TOKEN passes your OpenRouter key. ANTHROPIC_API_KEY must be set to an empty string, otherwise Claude Code tries to use it and the auth conflicts. CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC stops telemetry calls that would fail against OpenRouter.

That's verbose to type every time. I wrapped it in a shell function:

# ~/.zshrc
orclaude() {
  ANTHROPIC_BASE_URL="https://openrouter.ai/api" \
  ANTHROPIC_AUTH_TOKEN="$OPENROUTER_API_KEY" \
  ANTHROPIC_API_KEY="" \
  CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 \
  claude "$@"
}

Now orclaude launches Claude Code through OpenRouter. Regular claude still uses the Anthropic API directly.

You can also swap models. Claude Code has three model slots (Opus, Sonnet, Haiku) that you can override:

export ANTHROPIC_DEFAULT_SONNET_MODEL=\
  "google/gemini-2.5-pro"
export ANTHROPIC_DEFAULT_HAIKU_MODEL=\
  "google/gemini-2.5-flash"

Any model on OpenRouter that supports tool use works. Models without tool use will break Claude Code since it relies on tools for file reading, editing, and command execution.

OpenCode

OpenCode has native OpenRouter support.

What didn't work

I tried putting the API key in opencode.json. It passed config validation but OpenCode still tried to bill through its own payment system. "No payment method" error pointing at OpenCode's billing page, not OpenRouter's.

What worked

OpenCode picks up OPENROUTER_API_KEY from the environment automatically. The config file only needs the model names. opencode.json goes in the project root where you launch it, or the global ~/.config/opencode/opencode.json.

{
  "model": "openrouter/anthropic/claude-sonnet-4-6",
  "small_model": "openrouter/anthropic/claude-haiku-4-5"
}

The key has to be registered through OpenCode's auth system. Inside the TUI:

1. Run /connect
2. Search for "OpenRouter"
3. Paste your API key

After that, OpenCode routes through OpenRouter. The openrouter/ prefix on model names tells it which provider to use.

Note: OpenCode shows a warning when you select OpenRouter saying requests "will often be routed to subpar providers." That feels like they're trying to just sell their own model hosting.

OpenClaw

OpenClaw has native OpenRouter support. The model reference format is openrouter/<provider>/<model>, same as OpenCode.

In OpenClaw's home directory, openclaw.json, for the K8s setup /home/node/.openclaw/openclaw.json (the PVC mount). For a local install it'd be ~/.openclaw/openclaw.json.:

{
  env: { OPENROUTER_API_KEY: "sk-or-..." },
  agents: {
    defaults: {
      model: {
        primary: "openrouter/anthropic/claude-sonnet-4-6"
      },
    },
  },
}

Or use the CLI:

openclaw onboard \
  --auth-choice apiKey \
  --token-provider openrouter \
  --token "$OPENROUTER_API_KEY"

Since my OpenClaw runs in K8s with secrets injected via Vault, I added OPENROUTER_API_KEY to the Vault secret and updated the Vault Agent template to export it. Same pattern as the Linear API key from the previous post.

After redeploying, I cleared my OpenClaw agent's sessions (OpenClaw caches the model config per session) and asked what model it was running:

Claude: What model are you running on?
Pai: I'm running on Claude Sonnet 4.6 via OpenRouter
  (openrouter/anthropic/claude-sonnet-4-6).

Previously Pai ran Gemini 2.5 Flash through Google's API directly. Now it's Claude Sonnet 4.6 through OpenRouter. Same Vault-injected credentials pattern, different model, one billing dashboard.

The Linear skill still works through the new model. Pai listed my teams and issues the same as before. The skill's curl commands don't care which model is driving the agent.

Custom OpenRouter Usage MCP

With everything routed through OpenRouter, I wanted to check my spend without leaving the terminal. OpenRouter has a billing API, so I built a small MCP server that queries it.

The server has two tools. get_usage calls the /api/v1/key endpoint and returns daily, weekly, monthly, and total spend plus remaining credits. get_model_pricing calls /api/v1/models and shows per-million-token costs, with an optional filter to search by model name.

The whole thing is one TypeScript file using the MCP SDK with stdio transport. It reads OPENROUTER_API_KEY from the environment, same key the other tools already use.

Registering it

In ~/.claude.json, under the project's mcpServers:

"openrouter": {
  "type": "stdio",
  "command": "node",
  "args": [
    "/path/to/openrouter-mcp/build/index.js"
  ]
}

No env block needed. Claude Code spawns it as a child process, which inherits the shell environment.

Usage output

After restarting Claude Code, I asked it to check my OpenRouter usage:

Key: sk-or-v1-3fd...374
Free tier: no

Usage:
  Today:  $0.5467
  Week:   $0.5467
  Month:  $0.5467
  Total:  $0.5467

Credit limit:
  Limit:     $200.0000
  Remaining: $199.4533

That $0.55 is from a few hours of Claude Code and OpenCode use. The $200 limit is what I set on the key in OpenRouter's dashboard.

The pricing tool is useful for comparing models before switching. Searching "claude-sonnet-4" shows all the Sonnet variants and their per-million-token rates side by side.

Model pricing snapshot

Pulled from get_model_pricing on 2026-03-10:

Model	Prompt	Completion	Context
openai/gpt-5.4-pro	$30.00/M	$180.00/M	1050k
openai/gpt-5.4	$2.50/M	$15.00/M	1050k
openai/gpt-5.3-codex	$1.75/M	$14.00/M	400k
openai/gpt-5.2-pro	$21.00/M	$168.00/M	400k
openai/gpt-5.1	$1.25/M	$10.00/M	400k
anthropic/claude-opus-4.6	$5.00/M	$25.00/M	1000k
anthropic/claude-sonnet-4.6	$3.00/M	$15.00/M	1000k
anthropic/claude-haiku-4.5	$1.00/M	$5.00/M	200k
anthropic/claude-opus-4	$15.00/M	$75.00/M	200k
anthropic/claude-sonnet-4	$3.00/M	$15.00/M	200k
google/gemini-3.1-pro-preview	$2.00/M	$12.00/M	1049k
google/gemini-2.5-pro	$1.25/M	$10.00/M	1049k
google/gemini-2.5-flash	$0.30/M	$2.50/M	1049k
google/gemini-2.5-flash-lite	$0.10/M	$0.40/M	1049k
deepseek/deepseek-v3.2	$0.25/M	$0.40/M	164k
deepseek/deepseek-r1-0528	$0.45/M	$2.15/M	164k
meta-llama/llama-4-maverick	$0.15/M	$0.60/M	1049k
mistralai/mistral-large-2512	$0.50/M	$1.50/M	262k
qwen/qwen3.5-397b-a17b	$0.39/M	$2.34/M	262k
qwen/qwen3-coder-plus	$0.65/M	$3.25/M	1000k

Prices are per million tokens. These change often. Check OpenRouter's models page for current rates.