{"ok":true,"site":{"host":"ai-prices.oxweb.xyz","url":"https://ai-prices.oxweb.xyz/","type":"live","desc":"Compare normalized AI API prices across a curated set of public inference providers. WebMCP lets AI agents search models, compare providers and calculate workload costs from the same data shown on…","tools":[{"name":"search_ai_model_prices","kind":"answer","impl":"imperative","description":"Search the AI API Prices index for AI models and the discounted per-token prices that inference providers currently publish for them. Filter by model or family name, by the company that makes the model, by a maximum input price, by a maximum output price, and by how many providers sell it. All prices are USD per 1,000,000 tokens. Results are ordered cheapest first using the site ranking (input + 3 x output). Use this to answer \"which models cost under $X\" or \"find model Y\".","title":"Search AI model prices","annotations":{"readOnlyHint":true,"untrustedContentHint":true},"inputSchema":{"type":"object","properties":{"query":{"type":"string","maxLength":120,"description":"Model or family name, e.g. \"gpt 5.6 sol\", \"claude opus\", \"deepseek\"."},"maker":{"type":"string","maxLength":120,"description":"Company that makes the model, e.g. \"Anthropic\", \"OpenAI\", \"Google\"."},"max_input_usd_per_1m":{"type":"number","minimum":0,"description":"Only models with a provider charging at most this much per 1,000,000 input tokens, in USD."},"max_output_usd_per_1m":{"type":"number","minimum":0,"description":"Only models with a provider charging at most this much per 1,000,000 output tokens, in USD."},"min_providers":{"type":"integer","minimum":1,"description":"Only models sold by at least this many providers."},"limit":{"type":"integer","minimum":1,"maximum":25,"default":10,"description":"Maximum number of models to return."}},"required":[],"additionalProperties":false},"hasHandler":true,"hasExecute":true,"__source":"spec","executable":true,"handlerField":"execute","page":"/","pages":["/"]},{"name":"compare_ai_model_providers","kind":"answer","impl":"imperative","description":"List every inference provider selling one specific AI model, ranked cheapest first, with each provider's input price, output price, cache prices, saving against the model maker's official standard API price, and when the price was last observed. All prices are USD per 1,000,000 tokens. Use this to answer \"who is cheapest for model X\" or \"show me all providers for X\". Providers whose last refresh failed are excluded unless include_stale is true, and can never be reported as cheapest.","title":"Compare providers for one AI model","annotations":{"readOnlyHint":true,"untrustedContentHint":true},"inputSchema":{"type":"object","properties":{"model":{"type":"string","maxLength":120,"description":"The model to compare providers for, e.g. \"GPT-5.6 Sol\" or \"claude-opus-5\"."},"limit":{"type":"integer","minimum":1,"maximum":25,"default":10,"description":"Maximum number of providers to return."},"include_stale":{"type":"boolean","default":false,"description":"Include providers whose most recent price refresh failed. Their prices may be out of date."}},"required":["model"],"additionalProperties":false},"hasHandler":true,"hasExecute":true,"__source":"spec","executable":true,"handlerField":"execute","page":"/","pages":["/"]},{"name":"estimate_ai_workload_cost","kind":"act","impl":"imperative","description":"Calculate what a specific workload actually costs at each provider, given the number of input tokens and output tokens the job will use. Costs are computed from published prices in USD per 1,000,000 tokens and returned as exact USD totals per provider for every requested model, ranked cheapest first, naming the single cheapest model-and-provider combination overall. Pass raw token counts, not millions: 50 million input tokens is 50000000. Use this to answer \"which provider is cheapest for this workload\" or \"what would 50M in and 10M out cost\".","title":"Estimate the USD cost of a token workload","annotations":{"readOnlyHint":true,"untrustedContentHint":true},"inputSchema":{"type":"object","properties":{"models":{"type":"array","items":{"type":"string","maxLength":120},"minItems":1,"maxItems":5,"description":"One to five model names to price, e.g. [\"GPT-5.6 Sol\", \"Claude Opus 5\"]."},"input_tokens":{"type":"integer","minimum":0,"maximum":1000000000000,"description":"Total whole input (prompt) tokens for the whole workload, as a raw count. Fractional values are rounded."},"output_tokens":{"type":"integer","minimum":0,"maximum":1000000000000,"description":"Total whole output (completion) tokens for the whole workload, as a raw count. Fractional values are rounded."},"limit":{"type":"integer","minimum":1,"maximum":25,"default":5,"description":"Maximum number of providers to return per model."}},"required":["models","input_tokens","output_tokens"],"additionalProperties":false},"hasHandler":true,"hasExecute":true,"__source":"spec","executable":true,"handlerField":"execute","page":"/","pages":["/"]},{"name":"compare_ai_models","kind":"act","impl":"imperative","description":"Compare two to five AI models against each other on published discount-provider pricing: cheapest input price, cheapest output price, how many providers sell each, the best saving against the model maker's official price, and which provider is cheapest for each. All prices are USD per 1,000,000 tokens. Optionally pass input_tokens and output_tokens together to also compare the exact USD cost of a specific workload and get the cheapest model-and-provider overall. Use this to answer \"is model A or model B cheaper\".","title":"Compare two or more AI models","annotations":{"readOnlyHint":true,"untrustedContentHint":true},"inputSchema":{"type":"object","properties":{"models":{"type":"array","items":{"type":"string","maxLength":120},"minItems":2,"maxItems":5,"description":"Two to five model names to compare, e.g. [\"Claude Opus 5\", \"GPT-5.6 Sol\"]."},"input_tokens":{"type":"integer","minimum":0,"maximum":1000000000000,"description":"Optional. Total input tokens for a workload comparison. Must be given together with output_tokens."},"output_tokens":{"type":"integer","minimum":0,"maximum":1000000000000,"description":"Optional. Total output tokens for a workload comparison. Must be given together with input_tokens."},"limit":{"type":"integer","minimum":1,"maximum":25,"default":5,"description":"Maximum number of providers considered per model when costing a workload."}},"required":["models"],"additionalProperties":false},"hasHandler":true,"hasExecute":true,"__source":"spec","executable":true,"handlerField":"execute","page":"/","pages":["/"]},{"name":"show_ai_prices_in_page","kind":"act","impl":"imperative","description":"Update the AI API Prices page the person is currently viewing so it shows the result the conversation reached: sets the search box and switches between the Models and Providers views. The two views search different things. Models view shows a model's price comparison, including every provider selling that model — that is where the answer to \"which providers sell X, and which is cheapest\" belongs. Providers view is the provider directory and searches provider NAMES only, so a model name typed there matches nothing. Pass model to put that exact model's comparison on screen; when it resolves, the tool always opens Models view and ignores any view you asked for, because a model name cannot be found in the provider directory. Pass query on its own for a freer search, and use view: \"providers\" with it to browse the provider directory by provider name. Call this after answering a pricing question so the human can carry on browsing from that state instead of retyping the query. This only changes the existing on-page search and tab; it does not navigate anywhere.","title":"Show a result in the page the user is looking at","annotations":{"readOnlyHint":false,"untrustedContentHint":false},"inputSchema":{"type":"object","properties":{"model":{"type":"string","maxLength":120,"description":"A model to put on screen, e.g. \"GPT-5.6 Sol\". Resolved to its canonical name so the right comparison card is shown. Takes precedence over query."},"query":{"type":"string","maxLength":100,"description":"Text to put in the page search box, e.g. \"gpt 5.6 sol\". Pass an empty string to clear it."},"view":{"type":"string","enum":["models","providers"],"default":"models","description":"Which of the two existing views to show. \"models\" lists model price comparisons, including the providers selling a given model — use it for anything about a model, including its cheapest providers. \"providers\" is the provider directory and matches provider NAMES only, so use it only when browsing for a provider such as \"Surplus Intelligence\". Ignored when model is supplied and resolves: that always opens the models view."}},"required":[],"additionalProperties":false},"hasHandler":true,"hasExecute":true,"__source":"spec","executable":true,"handlerField":"execute","page":"/","pages":["/"]}],"category":"Developer Tools","favicon":"/favicons/ai-prices.oxweb.xyz.png","apiSurface":"spec","toolCount":5}}