# Cloud Pricing Calculators: full guide for agents Cloud Pricing Calculators (https://cloudpricingcalculators.com/) is a canvas-first front door to three pricing engines run by the same owner. Quotes are built by the engine of the cloud you pick; "All three" sends one request to all three engines; "Mirror on" rebuilds a priced quote on another cloud from a provider-neutral description of the workload. Prices are each provider's published prices as the engine found them; lines not priced live are flagged. ## When to use which Q: Which cloud is cheapest: Azure, AWS or Google Cloud? A: There is no single answer. The cheapest cloud depends on the workload: the services and sizes, the region, how many hours it runs, whether you commit for one or three years, which licences you need, and how much data leaves the cloud. The reliable way to find out is to price your own workload on all three. Describe it once here, pick "All three", and each cloud's own engine prices it side by side. Q: Which cloud pricing calculator should I use? A: If you already know your cloud, use that cloud's calculator: the Azure Price Calculator for Microsoft Azure, the AWS Price Calculator for Amazon Web Services, or the GCP Price Calculator for Google Cloud. If you have not decided, describe the workload here and price it on all three at once. Q: Where do the prices come from? A: Each quote is built by that cloud's own engine from its published price list. This site does not price anything itself and never translates one cloud's SKUs into another's. Prices that could not be confirmed live are flagged on every line. Q: Is it free? A: Yes. There is no paid plan. Each visitor has a free daily allowance of builds: 6 without signing in and 60 when signed in, where "All three" counts as three. The allowance resets at 00:00 UTC. Q: Do I need an account? A: No. A canvas is kept in your browser without signing in, and share links last 7 days. Sign in with your email to keep canvases in your account and make links that do not expire. ## AI links on this site Compose a link and give it to the user; it opens a live quote on the canvas. Use one provider block to price on one cloud, or two or three blocks to open a comparison group. BEGIN PARENT AI-LINK CONTRACT parent-1.0.0 (sister grammar 1.0.0) Grammar: https://cloudpricingcalculators.com/q///[///]{0,2}[?n=] := azure | aws | gcp (each at most once per link) := (-)* (the sister grammar, per provider) := (-)* (one spec per service, same order) := =(_=)* Rules: - One block prices on one cloud and lands one quote per service. Two or three blocks land a comparison group. - At most 3 services in total across all blocks, and at most 800 characters after /q/. - Lower case only; no "." and no "+". Unknown slugs and keys pass through to the engine. - If any part does not parse, nothing is priced (the link fails closed). - A sister link (https:///q//) also imports here; the cloud comes from its host. - The link carries requirements, never prices. The engines price it when the link is opened. Service slugs (Microsoft Azure, provider token "azure"): vm Azure Virtual Machines (also: vms, virtualmachines) sql Azure SQL Database (also: sqldb, azuresql) sqlmi Azure SQL Managed Instance postgres Azure Database for PostgreSQL (also: postgresql) mysql Azure Database for MySQL cosmos Azure Cosmos DB (also: cosmosdb) storage Azure Blob Storage (also: blob, blobstorage) files Azure Files disk Azure Managed Disks (also: disks) appservice Azure App Service (also: webapp, appservices) functions Azure Functions (also: func) aks Azure Kubernetes Service (also: kubernetes) containerapps Azure Container Apps aci Azure Container Instances redis Azure Cache for Redis vnet Azure Virtual Network loadbalancer Azure Load Balancer (also: lb) appgateway Azure Application Gateway (also: appgw) frontdoor Azure Front Door cdn Azure CDN firewall Azure Firewall vpngateway Azure VPN Gateway (also: vpn) bastion Azure Bastion keyvault Azure Key Vault monitor Azure Monitor (also: loganalytics) backup Azure Backup openai Azure OpenAI Service synapse Azure Synapse Analytics databricks Azure Databricks datafactory Azure Data Factory (also: adf) servicebus Azure Service Bus eventhubs Azure Event Hubs (also: eventhub) apim Azure API Management avd Azure Virtual Desktop search Azure AI Search (also: aisearch) staticwebapp Azure Static Web Apps (also: swa) signalr Azure SignalR Service logicapps Azure Logic Apps iothub Azure IoT Hub Service slugs (Amazon Web Services, provider token "aws"): ec2 Amazon EC2 rds Amazon RDS aurora Amazon Aurora dynamodb Amazon DynamoDB (also: dynamo) s3 Amazon S3 ebs Amazon EBS efs Amazon EFS lambda AWS Lambda ecs Amazon ECS eks Amazon EKS (also: kubernetes) fargate AWS Fargate elasticache Amazon ElastiCache (also: redis) redshift Amazon Redshift cloudfront Amazon CloudFront (also: cdn) elb Elastic Load Balancing (also: loadbalancer, alb) vpc Amazon VPC route53 Amazon Route 53 (also: dns) apigateway Amazon API Gateway (also: apigw) sqs Amazon SQS sns Amazon SNS cloudwatch Amazon CloudWatch lightsail Amazon Lightsail opensearch Amazon OpenSearch Service beanstalk AWS Elastic Beanstalk (also: elasticbeanstalk) Service slugs (Google Cloud, provider token "gcp"): gce Compute Engine (also: computeengine) gcs Cloud Storage (also: cloudstorage) cloudsql Cloud SQL bigquery BigQuery cloudrun Cloud Run cloudfunctions Cloud Functions (also: functions) gke Google Kubernetes Engine (also: kubernetes) cloudcdn Cloud CDN (also: cdn) loadbalancer Cloud Load Balancing (also: lb) spanner Cloud Spanner (also: cloudspanner) firestore Firestore bigtable Bigtable memorystore Memorystore (also: redis) pubsub Pub/Sub vertexai Vertex AI (also: vertex) persistentdisk Persistent Disk (also: pd) filestore Filestore clouddns Cloud DNS (also: dns) artifactregistry Artifact Registry cloudlogging Cloud Logging (also: logging) Spec keys (the same on every cloud; the engines read English): qty count of instances — e.g. qty=2 renders as "quantity 2" size instance or VM size — e.g. size=d4sv3 renders as "size D4s v3" vcores vCPU or vCore count — e.g. vcores=4 renders as "4 vCores" ram memory in GB — e.g. ram=16 renders as "16 GB RAM" nodes cluster nodes — e.g. nodes=3 renders as "3 nodes" os operating system — e.g. os=linux renders as "operating system Linux" region region code, no spaces — e.g. region=eastus renders as "region East US" tier service tier — e.g. tier=generalpurpose renders as "tier General Purpose" storage storage in GB — e.g. storage=250 renders as "250 GB storage" gb storage in GB — e.g. gb=500 renders as "500 GB storage" tb storage in TB — e.g. tb=2 renders as "2 TB storage" redundancy storage redundancy — e.g. redundancy=lrs renders as "LRS redundancy" hours hours per month — e.g. hours=730 renders as "730 hours per month" users expected users — e.g. users=500 renders as "500 users" requests requests per month — e.g. requests=10000000 renders as "10,000,000 requests per month" egress outbound GB per month — e.g. egress=500 renders as "500 GB outbound data per month" iops provisioned IOPS — e.g. iops=5000 renders as "iops 5000" throughput throughput hint — e.g. throughput=250 renders as "throughput 250" replicas replica count — e.g. replicas=2 renders as "replicas 2" env environment hint — e.g. env=prod renders as "env prod" commit commitment — e.g. commit=payg renders as "pay-as-you-go" Examples: https://cloudpricingcalculators.com/q/aws/ec2-rds/size=m5large_qty=2_region=useast1-engine=postgres_storage=100?n=Web%20app two AWS services on one link https://cloudpricingcalculators.com/q/azure/vm/size=d4sv5_os=linux/aws/ec2/size=m6ixlarge_os=linux/gcp/gce/size=n2standard4_os=linux?n=VM%20comparison a three-cloud comparison group, one service per cloud https://cloudpricingcalculators.com/q/azure/vm/size=d4sv5_qty=2_os=windows_region=eastus?n=Windows%20servers two Windows VMs on Azure in East US, with a name END PARENT AI-LINK CONTRACT parent-1.0.0 ## Federation API OpenAPI 3.1 document: https://cloudpricingcalculators.com/openapi.json. Builds return 202 and are polled; nothing is held open. Unsigned callers have a small free daily allowance. Keys, limits and the MCP server: https://cloudpricingcalculators.com/ai-agents. - POST /api/quote: Start a quote build on one engine, or on all three ("All three" fan-out). - GET /api/jobs/{jobId}: Job status. When complete, includes the normalised quote and details. - GET /api/groups/{groupId}: Status of a fan-out or mirror group, one entry per provider. - POST /api/groups/{groupId}/share: Create a parent share holding the group's finished nodes, laid out as a comparison group. - POST /api/mirror: Price the same workload on other clouds via the canonical workload description. - POST /api/canonical: Derive the canonical workload from a quote. A pure function; no spend, no limit charged. - POST /api/refresh: Re-price a node on the same provider as a new build (spine §4, Refresh price). - GET /api/import: Resolve a sister or parent share link or AI link into canvas content or a build plan. - GET /api/canvases: List the signed-in user's canvases ("My canvases"). - POST /api/canvases: Create a canvas. Also used to upload the browser canvas on first sign-in. - PUT /api/canvases/{canvasId}: Replace the canvas content (autosave, debounced 1.5 s, and Save). Optimistic concurrency via ETag. - POST /api/shares: Create an immutable share of a canvas snapshot. - GET /api/shares/{shareId}: Open a share, read-only. Never writes. - DELETE /api/shares/{shareId}: Revoke a share. Owner only; signed-in shares only, since unsigned shares have no owner. - GET /api/me: The signed-in user (created on first call for an SWA sign-in), and the sign-in methods on offer. - GET /api/limits: The caller's remaining builds today and per-provider availability. Not charged. - GET /api/health: Reachability of each sister engine and of the store, without spending any allowance. - GET /api/ai-link: The parent AI-link contract as data (ADR 0005), for agents such as the MCP compose_quote_link parent form. - POST /api/track: The web app's once-per-session visit beacon. Anonymous; always 204 for a well-formed body. - POST /api/converse: One turn of Guided mode's clarifying conversation (ADR 0011). Asks; never prices, never starts a build. - POST /api/auth/signup: Start email sign-up. Mails a single-use set-password link (24 h). Always the same 202. - POST /api/auth/forgot: "Forgot password". Mails a 1-hour reset link to an account's address. Always the same 202. - POST /api/auth/password: Set a password with an emailed token (first set or reset), and sign in. - POST /api/auth/login: Sign in with email and password. Every failure is the same 401. - POST /api/auth/logout: Sign out of an email session. The session is deleted server-side and the cookie cleared. - POST /api/auth/change-password: Change the signed-in email user's password. Revokes every other session. ## Sister calculators - Azure Price Calculator (Microsoft Azure): https://azure-price-calculator.com/ (agent guide https://azure-price-calculator.com/llms-full.txt, API https://azure-price-calculator.com/openapi.json) - AWS Price Calculator (Amazon Web Services): https://aws-pricing-calculator.com/ (agent guide https://aws-pricing-calculator.com/llms-full.txt, API https://aws-pricing-calculator.com/openapi.json) - GCP Price Calculator (Google Cloud): https://gcp-pricing-calculator.com/ (agent guide https://gcp-pricing-calculator.com/llms-full.txt, API https://gcp-pricing-calculator.com/openapi.json) - MCP server for all three clouds: https://mcp.azure-price-calculator.com/mcp (Streamable HTTP; tools build_cloud_quote, get_cloud_quote, compose_quote_link, list_calculator_capabilities) Estimates are informational. Not affiliated with Microsoft, Amazon or Google.