/
Product
A FinOps agent that watches,
investigates and follows through
Costory allocates your cloud, Kubernetes, data and AI bills, adds the tech context behind them, and runs a FinOps agent you work with from Claude Code or your assistant: it monitors spend around the clock, investigates changes, recommends optimizations and follows them up with the owning team in Slack.
First optimization recommendation
The next day
ROI
Pays for itself with the first recommendation you act on
Setup
About 30 minutes
Data residency
EU
Access
Read-only on your cloud
Always on
Monitors, detects, investigates, recommends, follows up, reports
01
Monitors
Each billing refresh is checked against team and service baselines, across every connected provider.
02
Detects
Anomalies and budget drift trigger alerts you can backtest before they go live.
03
Investigates
Cost changes are joined with deploys, owners, usage metrics and provider events.
04
Recommends
Concrete optimizations with evidence and an impact estimate from your own usage.
05
Follows up
It pings the owner, tracks the fix and confirms when spend is back to baseline.
06
Reports
Weekly showback per team and month-end packs, posted where each audience reads them.
Optimize
Recommendations grounded in your tech context
Costory gives the agent the context billing data lacks: deploys and releases, services and owners, usage metrics from CloudWatch, Cloud Monitoring or Datadog, and infrastructure metadata. That is what turns a cost line into a specific change, sent to the team that owns it.
Rightsizing from real utilization
Savings Plans, RIs and CUDs on your steady baseline
Idle and orphaned resources
LLM model and routing choices per feature
Storage tiers and lifecycle rules
Kubernetes requests and warehouse sizing
The agent recommends with the evidence attached. Your team decides what ships.
claude · checkout-api
>
Which databases in checkout-api are oversized for what they actually use?
●
costory - get_context (MCP)()
⎿
… +24 lines (ctrl+o to expand)
●
costory - query (MCP)(queries: [{type: "cost"}], filterCel: "cos_feature == 'checkout-api'", groupBy: "cos_service_name")
⎿
… +16 lines (ctrl+o to expand)
●
Rightsize the checkout primary from db.r6g.4xlarge to db.r6g.2xlarge.
- CPU p95 is 22% and memory is 41% over the last 30 days.
- Estimated impact: about €1,100/month at current usage.
Cost in context
Cost in context,
not just cost.
Connecting your bills is the easy part. Costory puts them in context: allocated with your telemetry, explained by your deploys, and documented for every follow up question.
01
Prepared
Allocated with your telemetry. Kubernetes usage per namespace, Datadog or Tsuga usage per service, BigQuery cost attributed to dbt models.
02
Proactive
Proposes the next move. Tags worth fixing first, shared cost to reallocate, a budget to create, an optimization to review.
03
Documented
Keeps a record of what happened. Past investigations, owners and decisions stay in your Context layer and inform every new answer.
Kubernetes allocation (EKS and GKE natively, AKS and other clusters through a Kubecost or OpenCost export) and dbt attribution for BigQuery are active on every plan, including Copilot, €250/month.
Context layer
The Context layer is everything Costory knows about how your company spends: how cost is allocated, who owns what, which metrics explain it, and what past investigations concluded.
Every assistant you connect starts from it, so a CFO in Claude and an engineer in Cursor get the same answer, and each investigation adds to it.
Why is our infrastructure cost per customer up 9% this quarter?
Two drivers, both owned and already investigated. The new EU region for Checkout adds about €14,000 a month since July (Platform team). In Support, the switch to a larger LLM in August doubled the cost per ticket; the team is testing a cheaper model for simple tickets.
Requests flat, token volume flat, cost climbing. A deploy put a timestamp at the top of the system prompt, and the cache hit rate dropped sharply. Costory flags it the same morning, with the deploy and the owner.
How it works
Four steps: connected in under 30 minutes, first recommendation the next day
01
Connect
Connect billing exports, install nothing
Set up native billing exports with Terraform or CloudFormation. Costory reads them and normalizes every provider into one cost model: clouds, Kubernetes, data platforms, observability and LLM APIs.
AWS CUR 2.0, Google Cloud billing export, Azure cost exports
Kubernetes, Snowflake, BigQuery, Datadog, OpenAI, Anthropic and more
Nothing installed in your infrastructure, no code deployed
AWS
Google Cloud
Kubernetes
Snowflake
OpenAI
02
Allocate
Allocate spend to teams, products and features
Virtual Dimensions map accounts, projects, labels and tags to the way your company is organised, even when tagging is incomplete. Costory suggests rules from your existing tags and provider metadata, and shared costs can be split by usage. Allocation of Kubernetes, BigQuery, Snowflake and other shared ressources are active on every plan, including Copilot, €250/month.
Allocation driven by your telemetry: Kubernetes usage, Datadog or Tsuga metrics, dbt models for BigQuery
AI-suggested allocation rules you can review and edit
Shared cost splitting, for example by Kubernetes usage
Unit costs: cost per customer, per request or per feature
03
Put it on call
The agent works in your Slack channels
Invite Costory to #finops or each team’s channel. When spend moves, the agent opens an investigation thread with the likely cause, pings the owner, proposes a fix and closes the thread once spend is back to baseline. Weekly showback and budget alerts run on their own.
Investigation threads with cause, owner and proposed fix
Weekly showback per team, in each team’s channel
Budget and anomaly alerts, backtested before they go live
Proactive suggestions: tags to fix first, shared cost to reallocate, a budget to create
finops
Costory
APP
9:00 AM
Weekly digest
Data platform is up 11% week over week (+€2,180). Main driver: BigQuery on-demand queries in the analytics project, starting after Tuesday’s deploy.
Open in Costory
Ask Claude
👀 2
✅ 3
2 replies
Daniel Chen
9:12 AM
@Costory
which BigQuery queries drove the increase?
Costory
APP
9:12 AM
On-demand queries in the analytics project, starting after Tuesday’s deploy. They account for the Data platform increase of +€2,180 (+11%).
AWS spend last week
+ ×
You
Why did our AWS bill go up last week?
Costory
›
Getting a context
›
Querying
↗ Explorer
›
Finding cost change factors
›
Suggesting actions
AWS spend rose 6.2% (+€3,140) last week.
1. Payments, EC2 in eu-west-1: most of the increase, after the autoscaling change deployed on Tuesday.
2. Search and Data platform: flat.
Ask a follow-up…
↑
04
Investigate and optimize
Your assistant does the deep work, with full cost context
Connect the Costory MCP to Claude Code, Claude, Cursor, Grok Bot or GitHub Copilot. The assistant queries allocated costs, lines up deploys and events, reads usage metrics and turns findings into recommendations, alerts and reports. Investigation notebooks keep the queries and conclusions together for review.
Hosted MCP server with OAuth login
Scoped to each user’s role and team
Investigation notebooks to share findings
Platform
Everything the agent, and your team, works with
01
Explorer and dashboards
Slice allocated cost by any dimension, compare periods and save dashboards for each team.
02
Unit economics
Combine cost with business metrics to track cost per customer, request or feature.
03
Anomaly alerts
Get alerted when a team, service or SKU moves outside its normal range, with the context to explain it.
04
Budgets
Set budgets by team or product and get warned before they are breached.
05
Events and deploys
Overlay deploys, provider events and custom events on cost so changes have a cause.
06
Investigation notebooks
Keep the questions, queries and findings of an investigation together and share them.
07
Sharing and exports
Share views by link, PNG or PDF, and export to BigQuery, Looker, Power BI or Pigment.
08
Access control
Domain auto-join on every plan; enterprise SSO on Copilot Pro. EU data residency.
09
Context layer
Costory prepares cost with your telemetry, proposes the next move, and keeps past investigations, owners and decisions in a Context layer every assistant starts from.
Clear limits
What Costory does not do
Costory does not buy Reserved Instances or Savings Plans for you. The agent recommends; your team decides.
It does not apply automated rightsizing or change resources for you.
It does not promise a savings percentage. Recommendations come with an impact estimate from your own usage.
It never gets write access to your cloud accounts. MCP actions create Costory objects only: alerts, reports, dashboards and events.
FAQ
Product questions
What is the Context layer?
The Context layer is everything Costory knows about how your company spends: how cost is allocated, who owns what, which metrics explain it, and what past investigations concluded. Every assistant you connect starts from it, so a CFO in Claude and an engineer in Cursor get the same answer, and each investigation adds to it.
Which allocation is included in Copilot, €250/month?
Allocation of Kubernetes, BigQuery, Snowflake and other shared ressources are active on every plan, including Copilot, €250/month, along with usage-based splits from Datadog, Tsuga and other connected metrics.
Does Costory change anything in our cloud?
No. Costory has read-only access to billing exports and the metrics you connect. It creates alerts, reports, dashboards and events in Costory. Rightsizing, commitments and other changes are decided and made by your team.
How fast do we get value?
Connecting your billing data and the Costory MCP takes under 30 minutes with Terraform or CloudFormation. AWS takes about 12 hours to deliver the first billing export, so your first optimization recommendation arrives the next day.
What does Costory’s agent do?
It watches allocated cloud, AI and data spend, investigates changes, recommends optimizations and follows them up in Slack. You work with it from the assistant you already use, such as Claude, Claude Code, Grok Bot or Dust. Costory provides the data, the context and the always-on monitoring; your team approves what goes live.
Is Costory right for us?
You can run FinOps with Claude and your provider’s MCP, and with one cloud, spend under €100k a year and one person on the bill, that is often enough. Costory makes it work as you grow: data across providers ingested and normalized, costs allocated to teams, an agent watching 24/7, and numbers you can verify down to the invoice line.
How is it different from asking Claude directly?
Claude answers when you ask, from the data you give it. Costory gives Claude ingested, normalized and allocated data you can verify, FinOps skills included automatically through the MCP, and an agent that keeps watching 24/7 and alerts the owner when spend moves.
Do I need a Costory account to use the Costory MCP?
Yes. The Costory MCP is hosted and works with your Costory account: it serves your organization’s allocated cost data and Context layer. Start with the 30-day free trial or the setup prompt.
Connect in under 30 minutes. Get your first optimization recommendation the next day.
30-day free trial. Copilot, €250/month, or Copilot Pro, €1,000/month, billed annually.
First optimization recommendation
The next day
ROI
Pays for itself with the first recommendation you act on
Setup
About 30 minutes
Data residency
EU
Access
Read-only on your cloud