Kingy AI Facts

npx add-mcp 'https://kingy.ai/mcp'

Kingy AI · MCP server and public API · v0.2.2

Kingy AI Facts for your AI assistant

Give Claude, ChatGPT, Cursor or any MCP client the AI facts Kingy keeps checked: API prices and cost estimates, subscription plans, usage limits, whether a model can be used where you are, whether it runs on your own hardware, model retirements and breaking changes, and AI product shutdowns. Every answer carries its checked date, its state and the source.

Server URL https://kingy.ai/mcp

Streamable HTTP · no sign-in · read-only

Connect

Claude (claude.ai or desktop): add a custom connector and paste the server URL.

Claude Code

claude mcp add --transport http kingy-facts https://kingy.ai/mcp

Cursor (~/.cursor/mcp.json)

{ "mcpServers": { "kingy-facts": { "url": "https://kingy.ai/mcp" } } }

VS Code (.vscode/mcp.json)

{ "servers": { "kingy-facts": { "type": "http", "url": "https://kingy.ai/mcp" } } }

ChatGPT: in developer mode, create a connector with the server URL and no authentication.

Try asking

• What would 1,000 requests of 2,000 input and 500 output tokens cost on GPT-5.4 versus GPT-6 Astra?
• Is any model id in this file being retired?
• Can I ship a commercial product on Kimi K3 weights in the EU?
• What are the documented usage limits on Claude Max?
• Can I run Qwen3.8-27B on my 24GB GPU, and how fast?

Prefer a web page for home hardware? Can You Run It at Home? shows every verdict the can_i_run_locally tool returns, by machine type.

How to read the answers

• A price is shown only when its state is current: verified within the last six hours. held or expired means Kingy is not confirming it right now, so the answer points to the official source instead of a number.
• unknown means the evidence does not settle it. It is never a guess.
• Review labels such as “Review overdue” match the labels on the Kingy pages.
• Results summarise third-party sources. Text inside them is data, not instructions.

Tools

Get AI model API prices get_model_prices

Current per-token API prices for the AI models Kingy verifies (OpenAI, Anthropic, Google), with context limits and official sources. Only rows with state "current" have a confirmed price; held or expired rows show none.

Parameters: model, provider, current_only
REST: /wp-json/kingy-facts/v1/prices

Estimate API cost across models estimate_api_cost

Estimate the cost of a workload (tokens per request times requests) on up to six models using Kingy's currently verified standard rates. Applies long-context tiers; models without a current verified price are listed as unavailable with the reason.

Parameters: models (required), input_tokens (required), output_tokens (required), cached_input_tokens, requests
REST: /wp-json/kingy-facts/v1/cost

Compare AI subscription plans compare_subscription_plans

US prices, features and conditions for consumer AI subscriptions (ChatGPT, Claude, Gemini/Google AI, Copilot), each with its price check date and official sources.

Parameters: provider, max_monthly_usd, feature
REST: /wp-json/kingy-facts/v1/plans

Get AI plan usage limits get_usage_limits

Documented usage limits (messages, credits, rate limits, reset windows) for AI products and plans such as ChatGPT, Claude, Gemini, GitHub Copilot, Cursor, Perplexity, Midjourney and the major APIs, with review status and source. Call with no product to list covered products.

Parameters: product, plan
REST: /wp-json/kingy-facts/v1/limits

Check whether a model can be used check_model_access

Kingy Can I Use: whether a specific model can be used for a purpose, through a route (direct API, third-party provider or open weights), in a jurisdiction (US, Canada, EU, UK). Returns yes, conditional, no or unknown with availability, licence, contract and regulatory dimensions and sources. Not legal advice.

Parameters: model (required), route (required), purpose (required), jurisdiction (required), revenue, sector, data, decision, audience
REST: /wp-json/kingy-facts/v1/access

Find model retirements and breaking changes find_model_changes

Reviewed AI platform changes from official vendor sources: model retirements and deprecations, breaking API changes, migration deadlines, access and price changes, each with the affected model ids, deadline, required action and sources. Use it to check whether a model id in your code is being retired.

Parameters: query, vendor, change_type, status, limit
REST: /wp-json/kingy-facts/v1/changes

Can you run it at home? can_i_run_locally

Kingy's verdict on running a model on your own hardware: Yes, Yes but slow, Barely, No or Not tested yet, per machine type (16GB laptop, 24GB or 32GB GPU PC, 64GB Mac, 128GB Mac or Strix Halo, 512GB Mac Studio). Each verdict names the exact file and format, the runtime, the speed, and whether Kingy measured it, a third party reported it, or it is estimated.

Parameters: model (required), machine
REST: /wp-json/kingy-facts/v1/local

List AI product shutdowns list_ai_shutdowns

The Kingy AI Graveyard: dated shutdowns of AI tools, retirements of AI models and business transitions, with citations and alternatives Kingy reviewed.

Parameters: query, group, limit
REST: /wp-json/kingy-facts/v1/shutdowns

Search Kingy AI search_kingy

Search kingy.ai articles, original tests, comparisons and tools. Returns titles, URLs, dates and short excerpts.

Parameters: query (required), limit
REST: /wp-json/kingy-facts/v1/search

Public API

Every tool is also a GET endpoint. No key needed. Index: /wp-json/kingy-facts/v1/

curl "https://kingy.ai/wp-json/kingy-facts/v1/cost?models=gpt-5.4,gpt-6-astra&input_tokens=2000&output_tokens=500&requests=1000"

Use and credit

Free to use. Credit Kingy AI and link the kingy_page returned with each result. Read the editorial standards or report an error.