This MCP server gives AI assistants read-only access to Compare Cloud Costs' weekly-refreshed cloud and AI pricing catalog through four tools.
Compare multi-cloud pricing for VMs, databases, serverless, storage, AI, GPUs, containers, and app hosting across AWS, Azure, GCP, Oracle, DigitalOcean, Alibaba, and Cloudflare, with filters for provider, vCPU, RAM, and region/geography.
Search AI LLM token pricing and context windows across providers such as OpenAI, Anthropic, DeepSeek, Mistral, Meta Llama, and Cohere, filtered by provider or model family.
Estimate end-to-end monthly costs for pre-modeled full-stack architectures using workload IDs, returning per-provider totals and component breakdowns.
Look up data center/region availability by country or city from catalog facets; compliance certifications are not actually exposed by the public API and should not be fabricated.
Requires a free CCC_API_KEY for tool calls; tools/list works without one, free keys are rate-limited, and pricing is refreshed about weekly rather than real-time.
Enables comparison of Cloudflare cloud pricing across compute, database, serverless, storage, AI, GPU, containers, and app-hosting categories.
Enables comparison of DigitalOcean cloud pricing across compute, database, serverless, storage, AI, GPU, containers, and app-hosting categories.
Enables comparison of Google Cloud pricing across compute, database, serverless, storage, AI, GPU, containers, and app-hosting categories.
Provides token pricing and context window data for Meta Llama models in AI model searches.
Provides token pricing (input, output, prompt caching) and context window data for OpenAI models in AI model searches.
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., "@Compare Cloud Costs MCP Server estimate monthly cost for a 3-tier web app on AWS vs Azure"
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.
A standalone Model Context Protocol server that gives your AI assistant access to the Compare Cloud Costs pricing catalog — cloud and AI model pricing across AWS, Azure, Google Cloud, Oracle, DigitalOcean, Alibaba, Cloudflare and the frontier AI providers, aligned to the FinOps FOCUS™ standard.
Data freshness: CCC refreshes its pricing information about once a week. The catalog is not real-time or live data, so use it for planning and comparison rather than minute-by-minute price checks.
This package is the installable, stdio-transport MCP server. Tool names and schemas match the canonical catalog used by the Compare Cloud Costs backend, and every tool is backed by the site's real public APIs — no sample data.
Tool | What it does |
| Query and compare multi-cloud compute, database, and serverless pricing across AWS, Azure, Google Cloud, Oracle, DigitalOcean, Alibaba, and Cloudflare. Filters: |
| Search token pricing (input, output, prompt caching) and context windows for AI LLMs across OpenAI, Anthropic, DeepSeek, Mistral, Meta Llama, and Cohere. Filters: |
| End-to-end multi-cloud monthly cost benchmarks for realistic full-stack architectures (VMs + managed DB + load balancers + storage + egress). Takes a |
| Checks global data center availability and region locations from the pricing catalog, with pointers to certification details. Filters: |
estimate_workload_architecture accepts the workload slugs from https://comparecloudcosts.com/workloads, e.g. 3-tier-web, serverless-web-app, rag-ai-app, k8s-app-platform, ecommerce-microservices, streaming-analytics, data-warehouse-bi, ml-training-hosting, agentic-ai-swarm. (The example slugs in the tool description — three-tier-web-app, ai-rag-pipeline, etc. — predate the current catalog; if an ID is unknown the tool returns the list of valid IDs.)
Related MCP server: CloudCost MCP Server
Tool calls need a free Compare Cloud Costs API key:
Go to https://comparecloudcosts.com/mcp and click Generate preview key (1-click, no card). Keys look like ccc_live_….
Set it as the CCC_API_KEY environment variable in your MCP client config (see below).
tools/list works without a key; tools/call returns an error explaining how to get one. Free preview keys are rate-limited per hour (the remaining quota is reported in tool responses).
Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json; Windows: %APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"compare-cloud-costs": {
"command": "npx",
"args": ["-y", "@comparecloudcosts/mcp-server"],
"env": {
"CCC_API_KEY": "ccc_live_your_key_here"
}
}
}
}Add to ~/.cursor/mcp.json (or a project .cursor/mcp.json):
{
"mcpServers": {
"compare-cloud-costs": {
"command": "npx",
"args": ["-y", "@comparecloudcosts/mcp-server"],
"env": {
"CCC_API_KEY": "ccc_live_your_key_here"
}
}
}
}Add to .vscode/mcp.json (note VS Code uses the servers key):
{
"servers": {
"compare-cloud-costs": {
"command": "npx",
"args": ["-y", "@comparecloudcosts/mcp-server"],
"env": {
"CCC_API_KEY": "ccc_live_your_key_here"
},
"type": "stdio"
}
}
}Requires Node.js ≥ 18.
All pricing data comes from GET https://comparecloudcosts.com/api/v1/pricing with header X-API-Key: <CCC_API_KEY> (documented in the "Public API" section of https://comparecloudcosts.com/docs). The endpoint is a faceted filter over the pricing database; the mappings below follow its server-side filter semantics exactly.
compare_cloud_pricingTool argument | API parameter | Notes |
|
| The API has no category filter that accepts these values — its |
|
| Provider slugs pass through unchanged ( |
|
| Numeric minimum. |
|
| Numeric minimum, GB. |
|
| The API has no region-slug parameter. Region IDs ( |
Results are ordered by price ascending (cheapest first) by the API. The server requests a 500-record page and returns the total count plus the 50 lowest-priced records, trimmed to the comparison-relevant fields.
search_ai_modelsproduct=ai plus provider (passed through) and model_family → the API's free-text search parameter. Records are returned with their attributes (modality, model tier, etc.); token prices are per the record's unit (typically per 1M tokens).
estimate_workload_architectureWired to the site's real estimator: POST https://comparecloudcosts.com/api/workloads with { "workloadId": "<id>" }, which returns per-provider monthly totals and a component breakdown (instance types, quantities, monthly prices) computed from catalog pricing records (refreshed about once a week). The tool returns those totals and components as-is. Unknown IDs return the API's "Workload not found" plus the list of valid IDs — numbers are never invented.
All 36 site architecture templates are supported:
Workload Name | Canonical Slug | Description |
Serverless Fallback & Resiliency Buffer |
| A resilient middleware buffer placed in front of serverless functions. Catches traffic spikes or cold-start timeouts by queuing requests into a persistent queue, caching session state, and dispatching to a failover VM worker pool. |
AI Gateway, Guardrails & Caching Proxy |
| An enterprise middleware proxy positioned between internal applications and external LLM vendors. Enforces prompt security guardrails (WAF), performs semantic prompt caching, rate-limits tenant requests, and manages API keys securely. |
Enterprise API Gateway & Auth Middleware |
| A standalone security & integration middleware layer placed in front of legacy or internal microservices to handle OAuth token validation, IP rate-limiting, TLS termination, and traffic shaping. |
Cross-Cloud Event Router & Message Broker |
| A cross-cloud middleware layer connecting disparate applications across AWS, Azure, GCP, and hybrid environments, converting event schemas and guaranteeing asynchronous delivery. |
RAG AI Application |
| A generative AI application using a Foundational Model connected to a Vector Database to search proprietary documents and return grounded answers. |
Containerized API via Load Balancer (ALB + Containers) |
| Exposes a containerized orders & payments API directly behind an internet-facing Application Load Balancer (Client → ALB → Managed Containers → Database) without an API Gateway. ALB handles path-based routing (/orders, /payments), target group health checks, traffic distribution, TLS termination, and native OIDC/JWT verification. |
Serverless HTTP API (HTTP Gateway + Functions + NoSQL) |
| A lean, low-cost serverless API pattern (Client → HTTP API Gateway → Serverless Functions → Managed NoSQL) featuring a native JWT authorizer, stage- and route-level throttling, and built-in CORS — with no load balancer or idle compute to provision. |
External Partner API (REST Gateway + Internal ALB + Containers) |
| A managed external partner API architecture (Partner → REST API Gateway → VPC Link V2 → Internal ALB → Managed Containers → Database). Combines REST API keys, per-client usage plans and quotas, request validation/transformation, and direct WAF integration at the edge with an internal ALB for private container health checks and traffic distribution. |
Serverless Web Application |
| A scalable, low-maintenance backend without provisioning servers. Perfect for event-driven web and mobile backends. |
Classic 3-Tier Web Architecture |
| The foundational blueprint for monolithic or traditionally scaled web applications using VMs and relational databases. The deliberate minimal baseline. |
Real-time Streaming Analytics |
| A highly demanded architecture for processing IoT telemetry, clickstreams, or financial data in real time. |
E-Commerce Microservices Stack |
| A resilient, decoupled architecture designed for high availability, fast product lookups, and fault tolerance. |
ML Training & Hosting |
| A pipeline for training machine-learning models on GPU clusters and serving them through a managed inference endpoint. |
Kubernetes App Platform |
| A general-purpose container platform for running microservices and stateful applications on managed Kubernetes. |
HPC / Scientific Computing |
| A high-performance computing cluster for simulations, modeling, and batch scientific workloads. |
SaaS on Managed Platform |
| A multi-tenant SaaS product running on a managed application platform (PaaS) with a relational backend. |
Data Warehouse & BI |
| A cloud data warehouse feeding business-intelligence dashboards, fed by an ETL pipeline. |
11-Layer Defense-in-Depth Production Architecture (Multi-AZ Analytics & Containers) |
| A production-grade 11-layer, Multi-AZ architecture demonstrating defense-in-depth from edge to data layer: Edge DDoS/DNS/TLS/GuardDuty/KMS, Layer-7 WAF, dual public ALBs (Shiny & RStudio), private ECS Fargate + EC2 Auto Scaling container fleets with outbound egress filtering, PrivateLink VPC isolation, SFTP (Transfer Family) + direct S3 ingestion, DataSync, shared EFS storage, and serverless SQL analytics (Athena via JDBC). |
Disaster Recovery (Warm Standby) |
| A cross-region warm-standby environment that can take over if the primary region fails. |
Content & Media Platform |
| A video/media platform that transcodes uploads and delivers content globally through a CDN. |
Compliance-Ready Web Application |
| A 3-tier web app hardened with managed security services for teams in regulated industries handling sensitive data. Security add-ons scale with the Security slider. |
RAG AI Knowledge Base |
| A retrieval-augmented generation pipeline for AI chat and search — document storage, a metadata store, an API layer, embeddings, orchestration, and a managed inference endpoint. |
Smart Manufacturing (IIoT) |
| An Industrial IoT (IIoT) platform for real-time sensor analytics — edge collection, stream ingestion, live metrics storage, historical analytics, and predictive maintenance. |
Event-Driven Image Processing |
| An event-driven architecture that processes uploaded images (e.g., ID verification, insurance claims) by triggering serverless functions to run AI vision models and store the extracted metadata. |
Hybrid Cloud Network Backbone |
| A robust networking architecture connecting on-premises data centers to cloud resources via dedicated connections and secure gateways. |
Zero-Trust Enterprise Edge |
| A secure, global entry point that strictly authenticates all users and devices, mitigating bots and edge threats before traffic hits the application. |
Event-Driven API Backend |
| An agile, fully serverless microservices backend combining managed API gateways, message queues, and rapid cache lookups. |
Voice AI Assistant |
| A speech-to-speech AI pipeline — incoming audio is transcribed, interpreted by a conversational LLM, and answered with a synthesized voice reply. Showcases how AI modality (Audio vs. Chat) and capability tier (Frontier vs. Standard) drive different AI product rows within one architecture. |
Autonomous Agent Swarm |
| A distributed multi-agent system where specialized AI agents autonomously collaborate, route tasks via message queues, and share transient state in memory. |
Distributed Saga & Compensating Workflows |
| Coordinates multi-service distributed transactions (e.g. Order → Payment → Inventory) using an orchestrator state machine. When a downstream step fails (such as stockout), automatically triggers compensating undo actions (payment refund, order cancellation) without distributed locks. |
Transactional Outbox & Stream Relay |
| Guarantees reliable dual writes between relational databases and event brokers. Orders and outbox events commit in a single ACID transaction, while a dedicated Change Data Capture (CDC) worker stream-relays messages to the event bus with zero message loss. |
Partitioned Pub/Sub with Consumer Groups |
| Scales event-driven task distribution across independent downstream systems (Email, Inventory, Analytics). Hash-keyed partitioning preserves strict per-entity ordering while competing consumer groups scale throughput horizontally with Redis idempotency guards. |
CQRS & Event Sourcing Architecture |
| Separates write commands from read queries for optimal scaling and zero-data-loss auditability. Changes are recorded to an append-only event store, and asynchronous projection stream workers continuously update read-optimized materialized views. |
CDC-Driven Cache Invalidation & Event Sync |
| Eliminates cache-aside race conditions, app-crash inconsistencies, and stale data by tailing database Write-Ahead Logs (WAL/binlog) via Change Data Capture (CDC). Stream events are consumed by an event-driven worker that evicts or refreshes affected keys directly in a distributed in-memory cache. |
Multi-Tier Layered Caching Architecture |
| A high-throughput, latency-optimized web stack employing a 3-tier caching hierarchy: Edge CDN (static assets & public HTTP caching), container in-process micro-cache (L1 with jittered TTLs), and a shared distributed cache (L2 Redis/Valkey cluster with stampede mutex locks) shielding transactional database read replicas. |
Customer Portal |
| A secure customer-facing portal — authenticated sign-in, edge protection, serverless APIs for spiky traffic, core services on Kubernetes, orders and payments with compensating workflows, managed relational + NoSQL data. |
Every slug accepted previously by the MCP server remains functional via automatic alias resolution:
streaming-analytics → mapped to realtime-streaming-analytics
three-tier-web-app & three-tier-web → mapped to 3-tier-web
ai-rag-pipeline → mapped to rag-ai-app
high-throughput-api → mapped to container-api-alb
batch-processing → mapped to hpc-scientific
lookup_datacenter_complianceRegional availability is computed from the site's facets endpoint, GET https://comparecloudcosts.com/api/facets?product=compute[&geography=…] (the same endpoint that powers the website's filter sidebar), which returns catalog-wide record counts per geography and per provider with proper faceted-search semantics. This is deliberately not derived from a sample of pricing records — the pricing endpoint returns records cheapest-first, which badly under-reports non-US geographies.
Honest limitation: per-datacenter compliance certifications (SOC 2, ISO 27001, HIPAA, PCI-DSS, FedRAMP, GDPR) are not exposed by any public API — the pricing records carry no certification fields. When a compliance_standard is requested, the tool says so explicitly and points to https://comparecloudcosts.com/datacenters and https://comparecloudcosts.com/compliance instead of fabricating coverage.
Also note: at the time of writing, the catalog's geography tagging for compute records is concentrated in North America (with a small Asia Pacific set), so geography-filtered queries for other regions may return zero records. That is the catalog's real coverage, reported as-is.
npm install
npm run build
npm start # stdio server on dist/index.js
node test/smoke.mjs # optional smoke test (set CCC_SMOKE_KEY to include API calls)server.json — official MCP Registry manifest (com.comparecloudcosts/mcp-server); the matching mcpName field is in package.json, as the Registry requires.
glama.json — Glama ownership-claim file (maintainer: the repository owner).
MIT
Query and compare multi-cloud compute, database, and serverless pricing across AWS, Azure, Google Cloud, Oracle, DigitalOcean, Alibaba, and Cloudflare.
Inputs: category (required), region, min_vcpu, provider, +1 more.
Calculates end-to-end multi-cloud cost benchmarks for realistic full-stack architectures (VMs + Managed DB + Load Balancers + Storage + Egress).
Inputs: workload_id (required).
Checks global data center availability, region locations, and compliance certifications (SOC 2, ISO 27001, HIPAA, PCI-DSS, FedRAMP).
Inputs: country_or_city, compliance_standard.
Search token pricing (Input, Output, Prompt Caching) and context windows for AI LLMs across OpenAI, Anthropic, DeepSeek, Mistral, Meta Llama, and Cohere.
Inputs: provider, model_family.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 4 tools
Each tool targets a distinct resource: AI model token pricing, cloud infrastructure pricing, full-stack workload estimation, and datacenter compliance. There is mild overlap between compare_cloud_pricing and estimate_workload_architecture since both deal with compute costs, but they operate at different levels of abstraction (raw pricing vs. composed architecture benchmarks) which keeps them separable.
All four tools follow a consistent snake_case verb_noun pattern (search_ai_models, compare_cloud_pricing, estimate_workload_architecture, lookup_datacenter_compliance). The verb choice is clear and predictable across the set.
Four tools is a lean but reasonable surface for a cost-comparison server, with each tool earning its place across distinct resource types. It is slightly thin—one might expect an additional tool for pure region/instance lookup—but nothing redundant is present.
The surface covers AI pricing, cloud pricing, workload estimation, and compliance/region lookup, which addresses the core cost-comparison workflow. Minor gaps exist, such as historical pricing trends or detailed per-instance configuration calculators, but agents can work around these.
Unified API to query AWS, GCP, Azure and generate Terraform/CLI execution kits for AI agents.
Verified cloud cost forecasting for AI agents. AWS, GCP, Azure pricing matrix.
Verified cloud cost forecasting for AI agents. AWS, GCP, Azure pricing matrix.
Cloud cost + FinOps knowledge for AI agents: AWS/Azure/GCP optimisation, AI spend, waste playbooks.
