Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
- Fixed model list in documentation: corrected "Meta-Llama-3.1-8B" to "Meta-Llama-3.1-70B" in the supported models table. - No other user-facing changes.
What This Skill Does
Pay-per-use AI gateway that routes requests to 19 models across Anthropic, OpenAI, Together.ai, Mistral, and Google via Bitcoin Lightning micropayments. Authentication works through a prepaid spend token or per-request Lightning invoice. No provider accounts or API keys required.
Eliminates the need to create accounts and manage separate billing with each AI provider.
When to use it
- Running AI inference without managing multiple provider API keys
- Comparing the same prompt across Claude, GPT-4, and Gemini in one tool
- Paying for one-off agent tasks with Bitcoin sats
- Using Pixtral vision or Codestral code models without a Mistral account
- Building autonomous agents that self-fund inference via Lightning payments
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: send a prompt to claude-opus-4-5-20251101 using a spend token
- 1Read LIGHTNINGPROX_SPEND_TOKEN from environment
- 2POST to lightningprox.com/v1/messages with model and message payload
- 3Set X-Spend-Token header for authentication
- 4Receive response and return model output
Claude's reply to the prompt, with cost deducted in satoshis from the prepaid token balance
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.