Every cloud, one ledger
Azure, AWS, Google Cloud and Open Telekom Cloud billing in one currency, one cost basis and one set of project names.
Know where your cloud budget goes
Self-hosted cost management for Azure, AWS, Google Cloud and Open Telekom Cloud. Built for the engineers who explain the bill and the finance teams who close the month.
Skip the live demo
Opening the live demo…Open full screen
Live demo with sample data. Click anything, or open it full screen.Your changes stay in this browser. Reset them in the demo sidebar, or open it full screen.
Reads billing from
Microsoft Azure
Amazon Web Services
Google Cloud
Open Telekom Cloud
From billing row to reviewed work. Costex brings every provider’s billing into one model, explains what changed, and turns savings into owned work your team can hand off.
Evidence before estimates
Every change, flag and saving keeps the rows, usage and assumptions behind it.
Ready for finance
Cost-centre assignments and native Excel workbooks that add up to the invoice.
Cost analysis
and anomalies
Compare any two periods, rank what moved and drill from project to resource. Unusual daily spend is flagged with the expected value beside the actual one.
Open Cost analysisSpend, last 30 days€391,669.59+€1,282.42 vs previous 30 days
- Azure €109,722
- AWS €102,924
- Google Cloud €97,130
- Telekom €81,894
Largest changes
- Data Platform productionGoogle Cloud · Managed Database +€698.03+€1,962.12
- Data Platform shared-servicesAWS · Object Storage +€580.66+€494.67
Features
Cost analysis
Period comparison
Compare any two periods across clouds, tenants and projects, and see exactly where the difference sits.
- Periods
- Calendar months, the last 7, 30 or 90 days, or a custom range.
- One basis
- Effective or billed cost, in one currency across every cloud.
Cost analysis
Drill-down
Open a project, then its services, then the resources that create the cost. Your filters stay as you go.
- Levels
- Cloud, project, service, region and resource.
- Charges
- The billing rows behind every figure stay one click away.
Cost analysis
Anomaly detection
Each day is compared with the same weekdays over up to eight weeks, trend included. A flag says what it costs and why it fired.
- Rule
- A day is flagged only when it clears the expected range by 50% and €100, so noise stays quiet.
- Credits
- Credits are excluded, so a discount never looks like a spike.
Cost analysis
Saved views
Save an investigation’s exact scope as a view your team can reopen.
- Dates
- Relative periods move with the calendar; an accounting month stays fixed.
- Access
- Everyone sees only the clouds and tenants they are allowed to.
Cost analysis
Weekly recap
A short deck of the week: spend and its yearly pace, the largest changes, open cost increases and savings work.
- Format
- Ten slides to read in a minute, from the Overview.
Cost analysis
Every cloud in one model
Azure, AWS, Google Cloud and Open Telekom Cloud billing in one model, compared in one currency.
- Sources
- Billing exports and provider APIs, refreshed by a background worker.
Savings
and handoff
Provider recommendations and Costex’s own checks land in one list, largest first. Each finding keeps its estimate, the usage behind it and an owner, and goes to the team as a Jira-ready draft.
Open OptimizeNeeds review 13
Stop ec2-portal-shared-01 outside working hoursCustomer Portal shared-servicesIdle resources€2,340
Idle resources€2,340Resize ec2-portal-prod-01 to m6i.xlargeCustomer Portal productionRightsizing€2,070
Rightsizing€2,070Stop ecs-dataplat-shared-01 outside working hoursData Platform shared-servicesIdle resources€1,935
Idle resources€1,9351-year commitment · s3.2xlarge.4 · ecs-portal-shared-01Customer Portal shared-servicesCommitment€1,800
Commitment€1,800Resize ecs-dataplat-prod-01 to s3.xlarge.4Data Platform productionRightsizing€1,665
Rightsizing€1,665Reviewing 1
1-year commitment · D8s_v5 · vm-dataplat-prod-01Data Platform productionCommitment€585
Commitment€585Planned 1
1-year commitment · m6i.2xlarge · ec2-dataplat-prod-01Data Platform productionCommitment€2,205
Commitment€2,205Implemented 1
Resize ec2-dataplat-shared-01 to m6i.xlargeData Platform shared-servicesRightsizing€2,475
Rightsizing€2,475Features
Savings
Recommendations
Provider recommendations and Costex’s own checks in one list, largest first.
- Providers
- AWS Cost Optimization Hub, Azure Advisor and Google Cloud recommenders.
- Own checks
- Utilization, idle resources and storage, from usage the clouds already report.
- Totals
- Findings for the same resource share an overlap group and count once.
Savings
Qualified estimates
Each estimate states how it was priced. The total leaves out anything that cannot be added safely.
- Billed rate
- Azure VM changes are priced at the rate the cost export billed for that size. Estimates at list price say so.
- Reserved sizes
- A resize of a size a reservation already pays for counts only the on-demand use it can replace. With none, it is held until renewal and left out of every total.
- Idle reservations
- Found in the Azure cost export, with when they idle and the use a reserved unit needs to pay off.
- Overlaps
- Alternatives for one resource, such as a resize and a reservation, count once.
Savings
Review workflow
Every finding moves from review to a verified result, with an owner and a next step at each stage.
- Ignored
- Ignored findings keep their reason and can be restored.
- Verified
- Measured savings are recorded after the change and kept apart from estimates.
Savings
Jira handoff
Prepare a Jira ticket for the right team: choose the board, confirm the cost centre and add a note.
- Draft
- The draft carries the evidence, resources, cost centre and potential savings.
- Sending
- Sending needs your Jira site connected. The demo shows a read-only preview.
Costex Agent
and its sources
Ask in plain language. The Agent reads your data through the same scoped tools as the app, cites every figure, and asks before it changes anything.
Learn about the AgentWhere should I start saving?
Start with Customer Portal shared-services. Stopping ec2-portal-shared-01 outside working hours is the largest open estimate, €2,340 per 30 days.
Idle resourcesStop ec2-portal-shared-01 outside working hours€2,340
Commitment1-year commitment · m6i.2xlarge · ec2-dataplat-prod-01€2,205
RightsizingResize ec2-portal-prod-01 to m6i.xlarge€2,070
Estimates can overlap. Check utilization and risk in Optimize first.
What changed in the last 30 days?
Spend rose €1,282.42 to €391,669.59. Data Platform production rose most, +€1,962.12, led by Managed Database.
Cost changesChange: EUR 1,282.42. 8 listed drivers. Basis: effective.
Create August’s exports for every tenant.
I prepared 8 PivotTable workbooks for August 2026, one per tenant, using your saved cost-centre assignments. Every project is assigned.
Create pivot exports
- Month
- August 2026
- Tenants
- 8
- Basis
- Effective cost
Nothing is created until you confirm.
Features
Costex Agent
Charts in answers
Ask what changed and the answer comes with the chart of what moved. A single figure or an explanation stays text.
- Scope
- The Agent reads only the clouds and tenants the person asking may see.
Costex Agent
Sources on every answer
Open any answer to see the query, period and basis behind each figure.
- Credentials
- Cloud credentials never reach the model.
Costex Agent
Confirmed actions
The Agent prepares changes as an exact preview. It cannot approve its own proposal; you confirm in the app.
- Examples
- Assign projects to cost centres, prepare a Jira handoff, create exports.
Budgets
and monthly close
Assign projects to cost centres once, track budgets and commitments against real usage, then give finance one workbook per tenant that adds up.
How monthly exports work| Tenant | Projects | Cost | Review |
|---|---|---|---|
| aws-production | 4 | €53,846.39 | Ready for draft |
| aws-shared-services | 4 | €52,036.75 | Ready for draft |
| azure-production | 4 | €57,855.69 | Ready for draft |
| azure-shared-services | 4 | €55,448.14 | Ready for draft |
| gcp-production | 4 | €50,276.78 | Ready for draft |
| gcp-shared-services | 4 | €48,167.68 | Ready for draft |
| otc-production | 4 | €43,188.54 | Ready for draft |
| otc-shared-services | 4 | €41,359.51 | Ready for draft |
Features
Monthly close
Cost centres
Map projects to cost centres with effective dates, and reuse the mapping every month.
- Corrections
- A mapping can start earlier without touching files already sent.
Monthly close
Excel exports
One native Excel workbook per tenant and month, with PivotTables by cost centre and every charge behind them.
- Drafts
- Exports stay drafts until provider billing is final.
Monthly close
Budgets and forecasts
See where the month will end. Each cloud is projected from its own last billed day, so a late export never looks like a drop.
- Projection
- Actual spend so far plus the mean of each cloud’s last 7 billed days for the days left.
- Targets
- Set a target from the projection, by cloud or project.
Monthly close
Commitments
Reservations, savings plans and committed-use discounts against what you actually use.
- Commitment usage is read from the billing export: credits against fees.