UptimeScout MCP server
OfficialRead-only access to regional service availability observations, history, statistics and disruption signals through seven MCP tools.
Search the catalog — find services and their availability observations (
uptimescout_services).Read one regional signal — status with observation time and explicit freshness (
uptimescout_status).Batch check — several service statuses in one call (
uptimescout_statuses).Summarize — reported availability across the catalog or selected services (
uptimescout_summary).Read history — weighted time-slot aggregates at 3-minute, 5-minute, hourly and daily resolutions (
uptimescout_history).Inspect statistics — weighted observation ratios from sample counts, not measured uptime (
uptimescout_stats).Find disruption signals — services with fresh elevated signals, not confirmed incidents (
uptimescout_incidents).Filter and page — by region (two-letter code), description language (20 codes), category, service IDs, history window (1–2160 hours), with
limit/offset.Safe by design — every tool is read-only, idempotent and non-destructive; no writes or side effects.
Allows checking regional availability observations for Cloudflare alongside other services, including current status signal, observation freshness, observation history, weighted statistics, and detection of fresh elevated disruption signals.
Allows querying regional availability observations for GitHub, including the current status signal, observation time and freshness, observation history at multiple resolutions, weighted statistics and summaries, and elevated-disruption incident signals.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@UptimeScout MCP server show cached status for api.stripe.com in eu-west and flag if it's stale"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
UptimeScout MCP Server
Connect AI assistants to regional service availability, observation history and disruption signals through seven read-only MCP tools.
Package: uptimescout-mcp · Quick start · Methods · Examples · Integration guide
Quick start
1. Get an API key. Compare the Free trial and paid plans. Licenses are sold and managed by Evelode. The same UptimeScout license works with REST, MCP and every UptimeScout SDK.
Set the key in your server environment:
export UPTIMESCOUT_LICENSE_KEY='your-license-key'PowerShell: $env:UPTIMESCOUT_LICENSE_KEY = 'your-license-key'.
2. Choose a transport.
Connection | Use it when | Setup |
Hosted Streamable HTTP | Your client supports a URL and Bearer authorization |
|
Local stdio bridge | Your client launches a command | Install the Python bridge below |
Shared jsonscraper MCP | You want several independently licensed products in one assistant connection |
For the dedicated hosted endpoint, configure the header Authorization: Bearer YOUR_EVELODE_LICENSE_KEY in your client's protected credential settings. No Python installation is needed. This endpoint uses Bearer authentication; the shared connector has its own OAuth setup.
Local stdio installation:
python -m pip install "git+https://github.com/jsonscraper/uptimescout-mcp-server.git@0.1.4"Then launch uptimescout-mcp from a client that inherits UPTIMESCOUT_LICENSE_KEY. Use the absolute executable path if your desktop client cannot find it.
Example stdio configuration for clients using an mcpServers object:
{
"mcpServers": {
"uptimescout": {
"command": "/absolute/path/to/uptimescout-mcp",
"env": {
"UPTIMESCOUT_LICENSE_KEY": "YOUR_EVELODE_LICENSE_KEY"
}
}
}
}The value above is a placeholder. Store your actual configuration privately. This JSON does not imply that all clients support environment-variable interpolation inside HTTP headers.
3. Ask your assistant: “Check GitHub and Cloudflare in GB. Report each signal, observation time and freshness. Do not describe another region as affected without checking it.”
Full setup, client configuration and troubleshooting · Example tool calls
Related MCP server: Context Layer Lab
Example response
A real public GitHub observation captured on 2026-11-18, shown as a historical example. It is not a live status claim. License-specific usage fields are omitted.
{
"status": "ok",
"region": "GB",
"service": {
"id": "github",
"name": "GitHub",
"category": "it",
"status": "operational",
"stale": false,
"observed_at": "2026-11-18T03:11:21.650479Z",
"age_seconds": 60,
"regions": [
"GB"
],
"signal": "reported_availability",
"scope": "regional",
"logo_url": "https://uptimescout.com/static/brand-assets/github.svg?v=18e8fe76f953a31c",
"logo_kind": "brand",
"description": "Platform for Git repositories, code collaboration, issues and software delivery.",
"description_language": "en"
},
"generated_at": "2026-11-18T03:12:21.928449Z",
"data_type": "availability_observations",
"stale_after_seconds": 600
}status at the top level describes the API request. service.status describes the availability signal. Always read stale, observed_at and region together. The region selects observations; lang independently selects service-description language.
MCP wraps the API result in result.structuredContent and also returns a text content block. Inspect result.isError before presenting a tool result as successful. Full MCP response envelope.
Main methods
Goal | REST endpoint | MCP tool |
Search the service catalog |
|
|
Read one regional signal |
|
|
Check several services in one call |
|
|
Summarize selected services |
|
|
Read observation history |
|
|
Inspect observation statistics |
|
|
Find fresh elevated signals |
|
|
Complete parameters and pagination · Response fields and errors · OpenAPI specification
Ready-to-run examples
Each data example performs one request. Pagination is explicit; no example downloads the entire catalog or history automatically. Run --help first to see usage without sending a data request. JSON files in the MCP repository are request bodies, not executable scripts.
Example commands for all seven methods
Regions, languages and logos
Use a two-letter country code such as GB, DE or JP; discover current coverage in available_regions. Coverage can differ by service. Without region, responses may aggregate observed countries, so specify it when you need a regional answer.
Descriptions support 20 language codes: en, ar, cs, de, es, fi, fr, he, it, ja, ko, nl, no, pl, pt, ru, sv, tr, uk, zh.
Service metadata includes description, description_language, logo_url and logo_kind. Use the versioned HTTPS logo URL with its HTTP cache headers. logo_kind distinguishes a brand asset from a monogram fallback.
Plans and request accounting
Plan | Requests per minute | Services per batch | History window |
Free | 60 | 5 | 7 days |
Starter | 300 | 10 | 7 days |
Builder | 600 | 25 | 30 days |
Pro | 1,800 | 50 | 90 days |
Scale | 3,600 | 100 | 90 days |
The rate window is 60 seconds. A successful data call counts as one request, including a batch within your plan's limit. Each additional history page is another call. Failed or rejected calls do not consume the successful-request allowance. Respect HTTP 429 and Retry-After when present.
Usage resets on license renewal or a plan change, not on the first day of a calendar month. Annual plans provide twelve monthly allowances as one pool for the full license term, with the same rate limit. History is available only for periods in which observations were collected.
Current prices and license-period allowances · License management
Understanding the data
Observations are updated every minute. A reported disruption signal is not a confirmed global outage or a direct uptime measurement. Regional signals can differ. Missing or stale observations remain unknown; sample statistics must not be labelled as an uptime SLA.
Configuration and troubleshooting
Hosted HTTP and local stdio configuration · Full integration guide
Support and license
The client source is MIT-licensed; hosted API access uses a separate UptimeScout license.
Service support · Report an SDK issue · Documentation
When reporting a problem, include the client version, method, region and error status. Redact license keys and authorization headers.
The SDK and bridge contain no installation or API-usage telemetry. Documentation links use campaign parameters for website referral analytics. Service names and logos identify monitored services and do not imply endorsement or partnership.
Remote MCP authorization
Connect to https://mcp.uptimescout.com with OAuth. Sign in through your Evelode account or enter a UptimeScout license key on the authorization screen. Secure Bearer headers and local stdio remain available. Connection options and revocation · Setup guide.
Available Tools
7 toolsuptimescout_historyBRead-onlyIdempotent
Read regional history: weighted time-slot aggregates at 3-minute, 5-minute, hourly and daily resolutions. Saved legacy aggregates keep their declared coarser resolution. Inspect resolution_seconds, samples, counts and effective_start; missing intervals are unknown.
Inputs: service (required), q, lang, hours, +5 more.
uptimescout_incidentsCRead-onlyIdempotent
List services with fresh elevated disruption signals; not confirmed provider incidents.
Inputs: q, lang, hours, limit, +5 more.
uptimescout_servicesCRead-onlyIdempotent
Search the service catalog and get service availability observations.
Inputs: q, lang, hours, limit, +5 more.
uptimescout_statsBRead-onlyIdempotent
Calculate weighted observation ratios from sample counts, including weighted time-slot aggregates. Not measured uptime. Inspect effective_start for UTC-slot alignment.
Inputs: service (required), q, lang, hours, +5 more.
uptimescout_statusCRead-onlyIdempotent
Get a service status with observation time and explicit freshness.
Inputs: service (required), q, lang, hours, +5 more.
uptimescout_statusesCRead-onlyIdempotent
Get a batch of service statuses.
Inputs: services (required), q, lang, hours, +5 more.
uptimescout_summaryCRead-onlyIdempotent
Summarize reported availability across the catalog or selected services.
Inputs: q, lang, hours, limit, +5 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 7 tools
The tools are mostly distinct: single status vs. batch status, catalog search, history, stats, summary, and incidents have different targeting. Some overlap remains between services/status and summary/stats because they all expose availability-related views, though the descriptions provide enough distinction.
All tool names share the same uptimescout_ prefix and use lowercase snake_case, which is predictable. The names are noun-based rather than action-based, and status/statuses are close but still intentionally distinguishable.
Seven tools is a well-scoped size for a read-only observability server covering catalog lookup, status retrieval, history, stats, summaries, and incidents. Each tool appears to earn its place without excessive surface area.
The toolset covers the core read-only lifecycle for uptime monitoring: service catalog, single and batch status, summaries, historical aggregates, statistics, and incident signals. Minor gaps may exist around additional catalog metadata or richer incident detail, but agents can work around them.
Maintenance
Related MCP Connectors
Read-only outage intelligence: live service status, report history, geography and spikes
Read-only MCP server for AIStatusDashboard status, incidents, metrics, and fallback recommendations.
Read public model catalog metadata with freshness timestamps. No model execution or payments.
Read-only cited financial datasets, bounded research queries and source diagnostics.
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables querying the current state, incident history, resolution times, and feed URLs for over 1,100 SaaS/cloud vendor status pages.7MIT
- AlicenseBqualityBmaintenanceRead-only MCP server that exposes governed operational records with provenance and validity windows so clients can check which evidence is still current before answering. It lets agents reconcile freshness, schedule state, and evidence specificity to determine whether scheduled automations are actually healthy, rather than trusting an expired dashboard summary.4MIT

senlayofficial
AlicenseNot gradedqualityNot gradedmaintenanceEnables MCP-capable clients to query verified physical context through location sensing, operability checks, and domain listing, while preserving evidence provenance and explicitly reporting unavailable inputs.Apache 2.0- FlicenseNot gradedqualityCmaintenanceEnables AI agents to query a locally cached, read-only view of a personal Zalo account, including connection status, conversations, bounded message history, cached text search, incremental events, and coverage, without sending, deleting, reacting, or marking messages read.-