If you have ever had to price out a cloud architecture before a client call, you already know the pain. You open the AWS Pricing Calculator, and then the real work begins. You add each service one at a time, hunt for the right instance type, fill in storage sizes, set the region, and triple-check that nothing got missed. A 10-minute task on paper turns into 45 minutes of clicking. Do that five times a week, and it adds up to hours that should have gone into actual architecture work instead.

That repetitive, error-prone process is exactly the gap a new, open-source AWS cost estimation tool is built to close. Instead of clicking through forms, you describe your infrastructure the way you would explain it to a colleague. The tool then hands back a real, working AWS Pricing Calculator link with the numbers already filled in.

Why Teams Need a Faster Way to Price AWS

Every cloud team runs into the same friction points when estimating AWS costs by hand:

• It is repetitive. The same service (EC2, RDS, S3, an ALB) gets re-added to the calculator over and over for every new proposal or sizing exercise.
• It is easy to mistype. A wrong storage value or the wrong instance type slips in unnoticed, so the final number is quietly off.
• It does not scale with speed. When a client or stakeholder asks “what would this cost,” waiting 30–45 minutes for a manual estimate kills momentum. The conversation moves on before the number is ready.
• It is hard to share collaboratively. Screenshots of a calculator page are not the same as a live, clickable link someone can open, adjust, and trust.

These pain points are precisely what pushed one solution architect to build a better AWS cost estimation tool.

The Architect Behind This AWS Cost Estimation Tool

Viresh Solanki built this tool after running into the same manual pricing grind described above, over and over, as part of his own day-to-day work. Viresh is a Solution Architect here on our team.

As a solution architect, Viresh regularly designs AWS infrastructure for clients and has to put a number on it before any of that infrastructure gets built. That means translating a proposed architecture into the AWS Pricing Calculator by hand, often multiple times a week, and often under time pressure when a client wants a ballpark figure quickly. Doing that manually for every revision of every proposal is exactly the kind of repetitive, time-consuming task that eats into the hours an architect should be spending on actual design work.

That firsthand frustration is the entire reason this AWS cost estimation tool exists. Rather than keep solving the same problem manually for each new engagement, Viresh built a tool that does the form-filling for him, then open-sourced it so other architects, DevOps engineers, and FinOps teams facing the same daily friction could use it too.

How the Tool Turns a Sentence Into a Pricing Link

The tool takes a sentence like this:

“2 t3.large EC2 with 50GB, RDS MySQL 100GB, 500GB S3, an ALB”

It returns an actual calculator.aws link with every one of those services added and priced, generated in seconds rather than minutes. There is no manual clicking. There is no need to hand-build a JSON payload either — plain English is the entire input.

Under the hood, this AWS cost estimation tool does the same thing a person would do by hand. It reads the request, then maps each phrase to the correct AWS service and configuration. Next, it drives a small browser engine to actually populate AWS’s own calculator and pull back the real, live cost. Because the numbers come directly from calculator.aws, they match what you would see if you built the estimate yourself, service by service.

Plain-English Input Makes Estimating Faster

The biggest time sink in manual estimation is translating a sentence in your head into calculator fields. For example, “a couple of medium EC2 instances, a database, some storage” has to become five separate form entries by hand. This tool removes that translation step entirely. You can describe:

• Instance types directly — “t4g.medium”, “db.m5.large”, “cache.r6g.large”
• Storage and transfer in plain units — “50GB storage”, “1TB data transfer”
• Request volume — “2M requests”, “10k messages per day”
• Snapshot schedules — “daily snapshot of 20GB” (weekly and monthly also work)
• Pricing models — “reserved 3 year”, “savings plan 1 year”
• Partial usage — “runs 5 hours per day” for workloads that are not always on

It currently recognizes roughly 50 AWS services. That list spans compute (EC2, Lambda, Fargate, EKS, Lightsail) and storage (S3, EBS, EFS, ECR). It also covers databases, including RDS across MySQL, PostgreSQL, Oracle, SQL Server, and MariaDB, plus Aurora, DynamoDB, Redshift, OpenSearch, and ElastiCache. Networking is covered too, from CloudFront, Route 53, and API Gateway to load balancers, VPC, NAT Gateway, Transit Gateway, VPN, Network Firewall, and PrivateLink. On the security side, it handles WAF, GuardDuty, KMS, Cognito, Inspector, and Security Hub, along with operational services like CloudWatch, CloudTrail, SQS, SNS, SES, Kinesis, Bedrock, and AWS Backup.

It is also forgiving about how you phrase things. Typos like “lamda,” “buckit,” or “dynmodb” still resolve correctly. Meanwhile, the tool tells you exactly what it understood before showing the final cost, so you can sanity-check the parsing rather than trust it blindly.

Solving the Collaboration Problem With a Real, Shareable Link

A second pain point manual estimation creates is shareability. A screenshot of a calculator session is a dead end, since nobody else can open it, adjust it, or verify the math. This AWS cost estimation tool solves that by always producing a real, official calculator.aws/#/estimate?id=... link rather than a static number. As a result, that link can be sent in Slack or pasted into a proposal. Anyone on the receiving end can open it and confirm the costs themselves, service by service, exactly as AWS’s own calculator would show them.

Three Ways to Use It, One Install

The tool is not locked into a single workflow. A single install (pipx install aws-calculator-mcp) sets up three different ways to use it:

1. As an MCP server for Claude, Cursor, Claude Code, Windsurf, and other AI-assisted IDEs. An assistant can design a stack from a plain description and generate the pricing link in the same conversation. No separate tool-switching is required.
2. As a CLI tool, either with a single prompt (aws-calc --prompt "...") or in an interactive mode. The interactive mode asks for region, estimate name, and infrastructure details step by step.
3. As an optional REST API for teams wiring cost estimation into ChatGPT Custom GPT Actions, n8n or Zapier automations, or any internal tool over HTTP.

Because all three share the same underlying parsing engine, a CLI estimate and an MCP-generated estimate for the same description return the same number.

Who This AWS Cost Estimation Tool Is Built For

The author’s stated goal in releasing this AWS cost estimation tool as open source was straightforward. Cost estimation on AWS should be transparent, shareable, and improved by the people who actually use it, rather than something every team quietly rebuilds in isolation. The intended audience is broad: a DevOps engineer sanity-checking a deployment, a solutions architect scoping a client proposal, a FinOps team auditing planned spend, or an independent builder pricing out a side project. The same plain-English workflow applies to all of them.

That commitment to transparency shows up in the documentation itself, since every example shown is a real, live calculator link rather than a static screenshot. Anyone evaluating the tool can therefore click through and verify the numbers rather than take a claim at face value.

Current Limits Worth Knowing About

This is an early release, and the author has been upfront that it is not a finished product. Recent updates have already added EBS snapshot support, automatic grouping of services by category (Compute, Database, Network, and so on), and tighter parsing around tricky cases like Aurora node counts and AWS Elastic Disaster Recovery configurations.

A few things are worth knowing before you rely on it for a final number:

• Very long, run-on prompts with many numbers close together (hours, disk counts, storage sizes, transfer amounts all in one sentence) can occasionally confuse the parser. Separating services with commas and attaching units to every number produces far more reliable results.
• A handful of calculator line items, notably AWS Backup and EC2 standard Reserved Instances, may still need a manual tweak after the link opens. For that reason, Savings Plans are recommended over standard Reserved Instances.
• The tool always shows what it understood from your prompt before reporting the cost. It is therefore worth a quick glance before trusting the total on anything important.

The author is actively looking for feedback on what is missing, what would make it genuinely useful for day-to-day work, and what edge cases still break it. That feedback goes directly into shaping the next release.

Try the AWS Cost Estimation Tool

For anyone who prices AWS infrastructure on a regular basis, this is worth a few minutes against a real architecture you already know the cost of, just to see how closely the result matches.

GitHub: github.com/vireshsolanki/aws-calculator-mcp PyPI: pypi.org/project/aws-calculator-mcp

The tool is MIT licensed, is not affiliated with AWS, and uses AWS’s own public calculator.aws endpoints to compute every cost it returns.