# WhatsDownNow > Live outage status for 250 popular services (AI model APIs, cloud and hosting, developer tools, payments, banks, ISPs, travel, streaming, social, gaming), plus outage alerts for the set of services a product depends on ("stack alerts"). Free JSON API, signed webhooks, and a remote MCP server. Verdicts combine each service's official status page with our own HTTP checks (about every 11 minutes) and anonymous user reports; they can lag or be wrong, so confirm critical decisions on the official status page. Not affiliated with any listed service. ## Set up outage alerts for a user (start here) If a user asks you to "set up WhatsDownNow alerts" for services they depend on, do this. No account, sign-up or API key is needed. 1. Get the services (common names are fine: "openai", "ChatGPT", "Anthropic", "AWS", "Google Cloud", "Stripe", "Chase") and where alerts should go: an email address (the user clicks one confirmation link) and/or a Slack, Discord or https webhook URL (active immediately). If they didn't say where, ask. Don't ask them to post their email publicly. 2. Create the stack, either way: - MCP (preferred): connect https://whatsdownnow.com/mcp (Streamable HTTP, no auth) and call create_alert_stack {"services":["openai","stripe","cloudflare"],"email":"user@example.com"} (or "webhook_url":"https://hooks.slack.com/..."). - REST: POST https://whatsdownnow.com/api/v1/stacks with JSON {"name":"My stack","services":["openai","stripe","cloudflare"],"emails":["user@example.com"]} (or "webhooks":["https://..."]). 3. The response includes id, token and manage_link. They are shown ONCE. Give the user the manage_link privately (it opens the stack in their browser, where they can send a test alert, reset the manage link, or delete it) and tell them to keep it private. Never post the token or manage_link publicly. If a manage_link may have leaked, call rotate_alert_stack_key (or POST /api/v1/stacks/{id}/rotate-key) with the current token; the old link stops working and a new one is returned once. 4. Tell the user to click the confirmation email if you used email. Optional: send_test_alert (MCP) or POST /api/v1/stacks/{id}/test to send a sample. Unknown service? The error lists did-you-mean slugs; or look up names with list_services {"query":"..."} (MCP) or GET https://whatsdownnow.com/api/v1/services?q=. If we don't track it, call request_service or POST https://whatsdownnow.com/api/v1/service-requests. Human-readable version with a copy-paste instruction: https://whatsdownnow.com/agents Example instruction a user might give you: "Set up WhatsDownNow outage alerts for OpenAI, Stripe and Cloudflare, sent to me@example.com. Connect the MCP server at https://whatsdownnow.com/mcp and call create_alert_stack (or, without MCP, follow whatsdownnow.com/llms.txt and use the REST API). Then give me the private manage link it returns." ## MCP server - Endpoint: https://whatsdownnow.com/mcp (Streamable HTTP, stateless, no auth). Client config: { "mcpServers": { "whatsdownnow": { "url": "https://whatsdownnow.com/mcp" } } } - Alert tools: create_alert_stack(services[], email?, webhook_url?, name?, on_down?, degraded_minutes?, reports_per_hour?, notify_recovery?), get_alert_stack(stack_id, token), update_alert_stack(stack_id, token, add_services?, remove_services?, add_email?, add_webhook_url?, name?), send_test_alert(stack_id, token), rotate_alert_stack_key(stack_id, token), delete_alert_stack(stack_id, token) - Status tools: get_status(slug), check_stack(slugs[]), list_services(query?, category?), list_incidents(slug), request_service(name, url?, reason?), get_ai_fallback(provider?), report_issue(service, error_type, model?, region?) - Readiness check before running a job: GET https://whatsdownnow.com/api/v1/status?slugs=chatgpt,claude,aws,stripe -> {"overall":"up|degraded|down|unknown","ready":true|false,...} - Service slugs: GET https://whatsdownnow.com/api/v1/services (e.g. chatgpt = OpenAI ChatGPT/API, claude = Anthropic Claude, gemini = Google Gemini, aws, google-cloud, cloudflare, vercel, supabase, stripe, github) ## API (OpenAPI 3.1: https://whatsdownnow.com/openapi.json) - GET /api/v1/services[?category=AI][&q=name] — monitored services (q finds a service by name) - GET /api/v1/status[?slugs=a,b][&category=] — verdicts for all services or a readiness summary for a set - GET /api/v1/status/{slug} — one service, with active official incidents - GET /api/v1/incidents/{slug} — recent official incidents + detected verdict changes - GET /api/v1/services/{slug}/history?days=7|30 — uptime %, incident count, p50/p95 of our HTTP probe (not model inference latency) - GET /api/v1/ai/health — AI API providers + recommended_fallbacks - GET /api/v1/ai/fallback?provider={slug} — healthy alternatives for one AI provider - POST /api/v1/reports — agent/client issue report {service|slug, error_type, model?, region?, detail?} - POST /api/v1/stacks — create alert stack; body {name, services: [slug | {slug,on_down,degraded_minutes,reports_per_hour}], defaults, notify_recovery, webhooks: [url], emails: [address], push_subscriptions: [PushSubscription]}; returns {id, token and manage_link (shown once), channels[].signing_secret} - GET | PATCH | DELETE /api/v1/stacks/{id} — with header "Authorization: Bearer " - POST /api/v1/stacks/{id}/test — send a test alert now - POST /api/v1/stacks/{id}/rotate-key — replace the management token and manage_link (shown once); the old ones stop working immediately ## Webhooks - JSON POST (payload_version 2) with event alert.triggered | alert.recovered | alert.test, service, verdict {old,new}, reason, severity, source (official|probe|reports|mixed), started_at, component, model, suggested_fallbacks (AI APIs), timestamps, status page links. v1 fields are unchanged. - Signature: X-WDN-Signature: t=,v1=hex(HMAC-SHA256(signing_secret, ".")) - Slack and Discord incoming-webhook URLs are auto-formatted. Retries with backoff (6 attempts); channel disabled after 15 consecutive failures. - Email channels require double opt-in; every email has one-click unsubscribe. ## Machine-readable - https://whatsdownnow.com/llms-full.txt — this file plus live status for every service - https://whatsdownnow.com/status/{slug}.md — Markdown version of each status page (current verdict, last check, recorded outages) - https://whatsdownnow.com/.well-known/mcp.json — MCP server manifest (endpoint + tool list) - https://whatsdownnow.com/openapi.json — OpenAPI 3.1 spec - https://whatsdownnow.com/sitemap.xml — every page ## Pages - https://whatsdownnow.com/ai — AI API provider board with live fallbacks and probe history - https://whatsdownnow.com/agents — set up alerts with your AI agent (copy-paste instruction, MCP connect snippets) - https://whatsdownnow.com/developers — human docs (API, clients, MCP setup for Cursor, Claude Desktop, Claude Code, VS Code) - https://whatsdownnow.com/stack — create alerts in the browser - https://whatsdownnow.com/status/{slug} — "Is X down right now?" page per service (e.g. https://whatsdownnow.com/status/chatgpt) - https://whatsdownnow.com/status/{slug}/history — outage history recorded by our checks - https://whatsdownnow.com/category/{category} — category hubs: ai, developer-cloud, productivity, google, microsoft, apple, social-streaming, money-shopping, games, home-security, banks-investing, internet-phone, travel-delivery