AI API Radar — live LLM API prices, status & price history

# ai-api-radar > Agent-payable live LLM API price/status feed — programmatically route to the cheapest operational model, read exclusive per-model price history, and subscribe to price-cut alerts. Built for autonomous AI agents. > Pay-per-call over the Model Context Protocol using the x402 "Payment Required" > standard (Coinbase x402, USDC on base). No signup, no API key to read the > free tiers; pay only for the premium result. ## MCP endpoint - https://aiapi.dropwatchhq.com/mcp (transport: streamable-http; call tools/list then tools/call) ## Paid tools (x402 pay-per-call, USDC on base) - `cheapest_model` — $0.01 USDC/call. Rank the cheapest currently-OPERATIONAL LLM API models meeting a task/context constraint, by live blended USD/1M-token cost across 11 providers (OpenAI, Anthropic, Gemini, Mistral, Cohere, Groq, DeepSeek, xAI, Together, Fireworks, OpenRouter). For an agent that wants to route each call to the cheapest model that still meets its context + capability needs, with provider status factored in. - `price_history` — $0.01 USDC/call. Return the full append-only $/M-token price time-series for one named LLM model — the only public per-model price HISTORY. For agents/analysts tracking when a provider cut or raised a model's price over time. ## Free tools - `status` — free. Current operational status for one or all AI API providers. - `price_change_alerts` — free. Subscribe a webhook to provider price changes, or read recent changes. ## How an agent pays (x402 over MCP) 1. POST https://aiapi.dropwatchhq.com/mcp with JSON-RPC tools/call for a paid tool. The result returns a FREE PREVIEW plus an `x402` descriptor inside structuredContent: { "x402Version":1, "accepts":[ { "scheme":"exact", "network":"base", "maxAmountRequired":"<atomic USDC>", "asset":"<USDC on Base>", "payTo":"0x38134a7F97636E26365Fb17405566595B917a341", "extra":{"name":"USD Coin","version":"2"} } ] } 2. Build an EIP-3009 transferWithAuthorization for maxAm