Open source FinOps control plane

Cloud cost visibility that reaches the infrastructure layer.

MaxOps gives cloud teams a local-first FinOps workflow for understanding infrastructure, finding waste, reviewing evidence, and acting with context without sending your cloud data to a hosted service.

Infrastructure Overview Local only
Scope Account, region, tags
View Potential action
Inform Inventory
Optimize Checks
Operate Actions
Resource Finding Next step
EC2 instance Idle capacity Inspect workload
EBS volume gp3 candidate Review conversion
S3 bucket Lifecycle gap Add retention policy

The cloud infrastructure problem

FinOps breaks down when cost signals are separated from operational reality.

Reports do not show resource context

Cloud bills and dashboards can identify spend, but engineers still need instance type, state, region, tags, workload, and service-specific posture before they can act.

Recommendations need evidence

Optimization work requires defensible checks, resource-level findings, current configuration, target action, and a way to exclude or snooze legitimate exceptions.

Actions happen in several tools

Teams move between dashboards, AWS Console, automation clients, and code. MaxOps keeps the same FinOps context available across those surfaces.

MaxOps workflow

Inform, Optimize, and Operate in one OSS stack.

MaxOps follows the practical FinOps lifecycle: understand what is running, identify waste with infrastructure checks, and run repeatable workflows with audit-friendly controls.

Inform

Build the infrastructure picture

Inventory views combine resource identity, service metadata, workload signals, tags, estimated cost, finding status, and region/account context.

Optimize

Turn checks into work

The backend check registry evaluates cloud waste patterns such as idle compute, unattached volumes, overprovisioned capacity, missing lifecycle policies, and old snapshots.

Operate

Keep decisions accountable

Snoozes, filters, check execution, action flows, and MCP write controls make optimization work explicit instead of hiding decisions in chat threads or spreadsheets.

Local-first by design

Your infrastructure data stays in your environment.

MaxOps is built to run locally: the backend, frontend, MCP server, and browser plugin work against your local MaxOps instance. Inventory, findings, tags, cost estimates, snoozes, and check results do not need to be sent to a third-party SaaS platform for the core workflow.

Local backend

The API runs in your environment and serves the dashboard, resource pages, check workflows, and MCP endpoints from your own MaxOps data store.

No external plugin calls

The browser plugin is scoped to AWS Console pages and local HTTP origins such as localhost or 127.0.0.1.

Private FinOps analysis

Teams can inspect infrastructure cost signals, workload evidence, and optimization findings without uploading internal cloud inventory to an external service.

Infrastructure visibility

From dashboard to resource type overview to instance-level detail.

MaxOps is designed for progressive investigation. Start with the operating picture, drill into service families, then inspect the exact resource with the evidence needed to make a change.

Dashboard

Cross-environment operating view

Group checks and findings by resource type, account, region, tags, and saved filters to see where optimization work should start.

Resource Type Overview

Service-specific inventory pages

Dedicated overviews for EC2, EBS, S3, RDS, DynamoDB, and ElastiCache compare health, workload, dimensions, findings, and potential actions.

Instance Details

Resource-level evidence

Detail pages expose configuration, tags, usage trends, MaxOps finding context, recommended action, cost estimate, and potential savings for the selected resource.

AWS Console context

Show MaxOps findings where operators already work.

The browser plugin overlays local MaxOps data inside AWS Console resource tables, adding cost, savings, and finding chips without calling third-party services.

NameMaxOps priceMaxOps findings
EC2 instanceCost chipIdle capacity
EBS volumeCost chipgp3 candidate

MCP workflows

Expose FinOps context to compatible clients.

The MCP server runs against your local MaxOps backend and provides tools for resource types, checks, latest results, inventory queries, tag cost summaries, snoozes, and check execution when write mode is enabled.

maxops://dashboard/summary query_inventory(resource_type="all") get_tag_cost_savings_summary(...)

OSS project stack

Separate components, one FinOps product surface.

API

Local API and check engine

FastAPI backend that runs locally for inventory, check execution, policy management, optimization APIs, onboarding, settings, scans, pricing, and MCP-facing endpoints.

UI

FinOps operating console

React + TypeScript operating console with dashboard, resource type overview pages, instance/resource detail pages, checks manager, notifications, security, settings, and onboarding.

MCP

MCP automation layer

Standalone MCP server that exposes inventory, checks, results, tag summaries, snooze workflows, and controlled write tools to desktop MCP clients.

Browser

AWS Console overlay

Dependency-free browser extension for Chrome, Edge, and Firefox that reads local MaxOps data only and overlays it in AWS Console resource views.

AWS optimization coverage

Checks across compute, storage, data, observability, and network cost patterns.

EC2 EBS S3 RDS DynamoDB ElastiCache ASG Aurora Lambda CloudWatch ECS EFS EMR Glue Kinesis OpenSearch Redshift Snapshots VPC

Run it locally

MaxOps is built as open-source infrastructure software.

Use the repository setup guide to run the backend, frontend, MCP server, and browser plugin locally. Keep your infrastructure data private while extending checks and workflows around your own FinOps operating model.