MCP-native · Claude, Cursor, Codex & more

Ask your AI anything
about your AWS bill.

Connect AWS, Azure, GCP, Datadog, Snowflake and 8 more in under 10 minutes. Query your entire cloud spend in plain English — no dashboards, no SQL, no exports.

Claude · nable connected
You
How much did we spend on compute across all cloud providers last month?
AI
Here's your compute breakdown for April 2025:

AWS EC2 + Fargate ········ $18,420
Azure Virtual Machines ·· $6,310
GCP Compute Engine ······ $4,870
Total ···················· $29,600

AWS is your largest provider at 62%. Spend is up 12% vs March — 3 new EC2 instances in us-east-1.
You
Any anomalies?
AI
● 1 high-severity anomaly — Datadog is up +127% vs your 28-day baseline ($4,100 vs avg $1,810). Spike started Tuesday. A Jira ticket has been filed automatically.

13

connectors out of the box

<10m

to full cost visibility

0

dashboards to maintain

Why not just call the APIs yourself?

Billing APIs give you data.
We give you answers.

You could spend a week writing adapters for 13 different auth models, schemas, and time formats. Or you could have this running in 10 minutes.

🔀

13 schemas, normalized to one

AWS, Azure, and GCP each return cost data in completely different formats, currencies, and granularities. We normalize everything so a single question spans all of them.

"Compare compute spend
 across all clouds this month"
↳ AWS $18,420 · Azure $6,310 · GCP $4,870
🧠

Intelligence the APIs don't have

Raw API data doesn't catch anomalies, suggest rightsizing, or file a Jira ticket at 2am. We layer seasonality-aware detection, RI analysis, and automated workflows on top.

● ANOMALY · HIGH
Datadog +127% vs same-weekday avg
Jira ticket filed · Slack alerted
💬

Your whole team can ask questions

Not just engineers. No SQL, no BI tool, no waiting for the data team. Anyone can ask a question and get a real answer in seconds — across every provider at once.

"Which team's spend grew the most this quarter?"
"Are we on track to hit our cloud budget?"
"What's our Snowflake spend per data pipeline?"

GETTING CONNECTED

Yes, every provider uses different credentials.
That's kind of the whole problem we solved.

AWS, Azure, and GCP all have completely different permission systems. Add Datadog, Snowflake, and eight more on top — and connecting your own data becomes a multi-week project. We did that work once, so you never have to.

🧭

A wizard that knows what to ask

Run finops setup and pick your providers. For each one, we tell you exactly what to create and what to paste — no prior cloud experience needed. Most teams are fully connected in under 20 minutes.

🔐

Your credentials never leave your machine

Everything you enter is encrypted and stored in your operating system's secure keychain — the same place your laptop stores passwords. We never see your keys. There's no cloud sync, no account to breach.

Done once. Works forever.

After setup, your AI client connects to a single local server that handles all 13 providers. No dashboards to maintain. No exports to schedule. Ask about AWS and Datadog in the same question — we normalize everything behind the scenes.

WHAT SETUP ACTUALLY LOOKS LIKE

1

Install

pip install finops-mcp

2

Run the wizard

Pick your providers. Follow the prompts. ~15 min total.

3

Add to your AI client

One config line in Claude Desktop, Cursor, or Windsurf.

Start asking questions

Every provider, one conversation. No exports needed.

Built for real FinOps

Not just data retrieval.
Actual intelligence.

🔴

Anomaly detection

Seasonality-aware z-score analysis — compares the same weekday vs prior weeks, not a flat rolling mean. Flags real spikes before they hit the invoice.

ANOMALY · HIGH
Datadog +127% vs same-weekday baseline
$4,100 today · avg $1,810 · z=3.8
🏷️

Tag-based cost attribution

Map cloud resource tags to teams, services, and environments via YAML rules. Ask "which team is burning the most?" and get an actual answer.

team: platform · $12,400/mo
team: growth  · $8,200/mo
team: data   · $6,800/mo

Rightsizing recommendations

Identifies EC2 instances running at low CPU over 14 days. Calculates exact savings from downsizing. Auto-files a Jira/Linear/GitHub ticket for each one.

i-0a3b8f · m5.2xlarge → m5.xlarge
avg CPU 8% · max 22% · save $192/mo
🔔

Proactive Slack & Teams alerts

Daily digest and anomaly alerts fire from the scheduler — no Claude session required. Weekly email digest lands in your inbox every Monday.

Slack · Teams · Email
Daily digest · Anomaly push · Weekly report
📊

RI & Savings Plan analysis

Real CE API data for coverage, utilization, and waste. Recommends Compute Savings Plan purchases with projected ROI and breakeven timeline.

📧

Invoice email parsing

Connects to your billing inbox via IMAP. Parses PDFs and HTML invoices from PagerDuty, New Relic, GitHub Enterprise — vendors with no billing API.

🔒

Credentials never leave your machine.

Every SaaS FinOps tool asks you to hand your cloud credentials to a vendor. We don't. nable stores all credentials encrypted with Fernet in your OS keyring — AWS, Azure, GCP, every SaaS key. The vault key lives on your hardware, not our servers.

Fernet encryption OS keyring storage Audit log Key rotation OAuth flows

13 connectors

Every cost source, one interface.

Click any connector for setup instructions.

AWS

Cost Explorer

Azure

Cost Management

GCP

Cloud Billing / BigQuery

Pricing

Simple, honest pricing.

One flat fee. All features. Cancel any time.

Enterprise

Custom

Volume discounts · Dedicated support

  • Everything in Pro
  • Unlimited connectors & history
  • Custom connector development
  • SSO / SAML
  • Dedicated Slack channel
  • SLA + 99.9% uptime guarantee
  • On-prem / private VPC deployment
Contact sales →
Payments secured by Stripe

Ready to stop guessing
what your cloud costs?

Set up in 10 minutes. Ask Claude anything.
14-day free trial, no credit card required.

14-day trial · No credit card · Cancel anytime