llms.txt
Compare with other sites' llms.txt files →# Costory
> Costory gives every engineer the cost context of a FinOps team. It is built for platform teams at companies with 40+ engineers running several cloud, data-cloud and LLM providers. Costory normalizes every bill (AWS, Google Cloud, Azure, Kubernetes, Snowflake, Databricks, BigQuery, Datadog, OpenAI, Anthropic and more) into one cost model on BigQuery, allocates it with rules versioned in git (YAML enrichments, reviewed in a PR, backfilled over full history, managed with Terraform) and with telemetry (Kubernetes, Datadog, Tsuga, dbt), and joins it with deploys and usage metrics. On top of that context runs the FinOps harness: an agent with skills, company context and loops (Scouts, agents assigned to reports and alerts with plain-text instructions) that monitors, investigates, recommends and follows up in Claude, Claude Code, Cursor, Slack, Dust and Backstage, through a hosted MCP server. Costory is read-only on your cloud and never touches your code base: it hands each recommendation to your coding agent or team. Pricing is flat and public: from €250/month (about $300), never a percentage of your bill. Plans: Starter €250/month, Team €600/month, Business €1,200/month (billed annually), Custom from €2,500/month. Starter fits the usage of 80% of customers. 30-day free trial, self-serve, no credit card required. Data hosted in the EU.
Comparison pages state the date their claims about other tools were last checked; those claims come from each vendor's public documentation and pricing pages. Costory is the publisher of this site.
## About
- [About Costory for AI and LLMs](https://www.costory.io/ai-info): Structured facts: what Costory is, who it is for, pricing, integrations, MCP tools, what it does not do, FAQ. Last reviewed November 2026.
- [Costory home](https://www.costory.io/): Cost context for platform teams on several clouds and LLM providers. Allocation as YAML in git, answers in Claude, Slack and Backstage. From €250/month, never a % of your bill.
- [About Costory](https://www.costory.io/about): Costory SAS, Station F, Paris. Built by data engineers; the cost context layer for engineering teams and their AI assistants. We run Costory on Costory.
- [Who Costory is not for](https://www.costory.io/not-for-you): If one person looks at the cloud bill, connect Claude to your billing API instead. Costory is for organizations where several teams, their leads and finance need the same number.
- [Build vs buy](https://www.costory.io/build-vs-buy): One dashboard with Claude is free. A shared internal FinOps tool takes weeks to build, then hosting, maintenance and feature requests forever. What it takes, and what you keep control of with Costory.
- [Moving to Costory from another FinOps tool](https://www.costory.io/switch): Try Costory next to the tool you have, free for 30 days, no credit card. Costory reads billing exports read-only from each provider, so it needs nothing from your current tool. Setup with Terraform, CloudFormation or by hand in under 30 minutes. Costory doesn’t import rules from other tools: ask Claude to rewrite them as enrichment YAML, then review the diff in a PR. From €250/month, never a % of your bill.
- [How Costory works](https://www.costory.io/product): One cost model across providers, allocated by YAML rules in git and by telemetry, measured with unit economics formulas and backtested alerts, and run by the FinOps harness in Claude, Slack and Backstage.
- [The FinOps harness](https://www.costory.io/finops-harness): Why cost has to come to you: Costory watches allocated cloud, data and LLM cost every day, pings the owner only when it is worth it, explains the cause, hands the fix to your coding agent and follows up until spend is back to baseline. Loops are Scouts, agents you assign to reports and alerts. Read-only.
## Product
- [FinOps MCP](https://www.costory.io/finops-mcp): Hosted MCP that gives Claude, Claude Code, Cursor, GitHub Copilot, Codex, Dust and any MCP client allocated cost across clouds, Kubernetes, data clouds and LLM APIs, with the deploys, owners and metrics that explain it. Read-only on your cloud, EU-hosted.
- [Costory for Slack](https://www.costory.io/slack): Scouts post cost alerts in each team's channel with the likely cause, the deploy and the owner, follow up until spend is back to baseline, and answer follow-ups with @Costory. Weekly showback and budget alerts included.
- [Costory in Claude](https://www.costory.io/assistants/claude): Use the Costory MCP from Claude to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Claude Code](https://www.costory.io/assistants/claude-code): Use the Costory MCP from Claude Code to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Cursor](https://www.costory.io/assistants/cursor): Use the Costory MCP from Cursor to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in GitHub Copilot](https://www.costory.io/assistants/github-copilot): Use the Costory MCP from GitHub Copilot to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in ChatGPT](https://www.costory.io/assistants/chatgpt): Use the Costory MCP from ChatGPT to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Codex](https://www.costory.io/assistants/codex): Use the Costory MCP from Codex to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Gemini CLI](https://www.costory.io/assistants/gemini-cli): Use the Costory MCP from Gemini CLI to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Grok Bot](https://www.costory.io/assistants/grok-bot): Use the Costory MCP from Grok Bot to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Dust](https://www.costory.io/assistants/dust): Use the Costory MCP from Dust to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Costory in Windsurf](https://www.costory.io/assistants/windsurf): Use the Costory MCP from Windsurf to query allocated AWS, GCP, Azure, Kubernetes and LLM cost, and to set alerts and reports.
- [Connect guides](https://www.costory.io/connect): Guides to connect Claude, Claude Code, Cursor, GitHub Copilot, ChatGPT and Codex to AWS, Google Cloud, Azure, Kubernetes, Snowflake, Datadog and LLM spend via the Costory MCP.
## Pricing
- [Pricing](https://www.costory.io/pricing): Flat, public pricing, never a percentage of your bill. Plans follow history, refresh frequency and usage; your own agent connected to the Costory MCP costs nothing extra and doesn't use Ask Costory credits. Spend bands are only a guardrail for very large estates. Starter €250/month and Team €600/month (up to €10M annual cloud and AI spend), Business €1,200/month (up to €20M), billed annually; Custom from €2,500/month above €20M. Unlimited users and data sources on every plan. Starter fits 80% of customers. 30-day free trial, no credit card required, self-serve.
## Comparisons
- [All comparisons](https://www.costory.io/compare): Costory compared with the AWS, Azure and Google Cloud MCP servers and with Vantage, CloudZero, Finout and Kubecost. Claims checked October 2026.
- [Cloud cost management tools in 2026: what each one does](https://www.costory.io/cloud-cost-management-tools): What each cloud cost management tool does: native tools, allocation platforms, Kubernetes, commitments and AI spend. Features to look for, with sources.
- [FinOps tools in 2026: how to choose by need](https://www.costory.io/best-finops-tools-2026): Twelve FinOps tools compared by fit, pricing model, assistant support and limits, with sources. Costory is our product. Last updated 15 October 2026.
- [AWS Billing MCP vs Costory](https://www.costory.io/compare/aws-billing-mcp-vs-costory): AWS's Billing and Cost Management MCP runs locally on your AWS credentials and covers AWS only; when it is enough and when a hosted layer across providers is useful.
- [What is a FinOps agent? AWS FinOps Agent and Costory, compared](https://www.costory.io/finops-agent): A FinOps agent monitors, optimizes and explains cloud spend. How AWS FinOps Agent (preview) and Costory differ: providers, where they work and fit.
- [Google Cloud Billing MCP vs Costory](https://www.costory.io/compare/google-cloud-billing-mcp-vs-costory): Google's official Cloud Billing MCP manages billing accounts, IAM and SKU prices but returns no spend data; how to ask an assistant about GCP spend.
- [Azure Cost Management MCP vs Costory](https://www.costory.io/compare/azure-cost-management-mcp-vs-costory): Microsoft's Azure MCP server adds Cost Management and Pricing toolsets; what it covers and where Azure-only coverage stops.
- [Costory vs AWS Cost Explorer and spreadsheets](https://www.costory.io/compare/cost-explorer): Costory vs AWS Cost Explorer and spreadsheets: when Cost Explorer is enough, and when automated reporting and explained changes across providers help.
- [Costory vs asking ChatGPT or a DIY MCP](https://www.costory.io/compare/chatgpt): Costory vs asking ChatGPT or a DIY MCP: when a DIY setup works, and why normalized, allocated data and FinOps tools give consistent answers.
- [Costory vs legacy FinOps platforms](https://www.costory.io/compare/legacy-tools): Costory vs legacy FinOps platforms: built for engineering teams without a FinOps department. Flat pricing. Connect in under 30 minutes, get your first optimization recommendation the next day.
- [Vantage vs Costory](https://www.costory.io/compare/costory-vs-vantage): Vantage and Costory compared on coverage, assistant access (Slack, Claude, Cursor) and setup, with Costory's flat pricing and 30-day free trial (no credit card required). Claims checked October 2026.
- [CloudZero vs Costory](https://www.costory.io/compare/costory-vs-cloudzero): CloudZero vs Costory: unit cost focus, MCP access model, pricing (quote vs published flat), setup time. When each is the right pick.
- [Finout vs Costory](https://www.costory.io/compare/costory-vs-finout): Finout's MCP is hosted and read-only; Costory's can create alerts, reports and dashboards. Coverage, pricing and fit compared.
- [Datadog Cloud Cost Management vs Costory](https://www.costory.io/compare/costory-vs-datadog-cloud-cost-management): Datadog CCM puts cost next to observability; Costory adds cloud, data and LLM providers and an MCP built for cost questions. Where each fits.
- [Kubecost vs Costory](https://www.costory.io/compare/costory-vs-kubecost): Kubecost (IBM) allocates in-cluster cost; Costory ingests Kubecost or OpenCost exports and adds cloud, data and LLM cost. When you need one or both.
- [nOps vs Costory](https://www.costory.io/compare/costory-vs-nops): nOps focuses on AWS commitment management and savings; Costory on allocation and context for engineers and their assistants. Which problem you are solving.
- [DoiT vs Costory](https://www.costory.io/compare/costory-vs-doit): DoiT bundles cloud resale, support and FinOps tooling; Costory is a FinOps harness: data, skills and checks your agents run on. How the models differ.
- [IBM Cloudability vs Costory](https://www.costory.io/compare/costory-vs-cloudability): Cloudability is built for a central FinOps team; Costory for platform engineers and their assistants. Setup, pricing model and fit compared.
- [Ternary vs Costory](https://www.costory.io/compare/costory-vs-ternary): Ternary vs Costory: a finance-owned technology ledger compared with a flat-priced cost context layer for engineers and AI assistants.
- [Flexera (Spot) vs Costory](https://www.costory.io/compare/costory-vs-flexera-spot): Flexera One FinOps and Spot compared with Costory: a broad optimization portfolio versus a flat-priced allocation layer for engineering teams.
- [Harness Cloud Cost Management vs Costory](https://www.costory.io/compare/costory-vs-harness-ccm): Harness Cloud Cost Management compared with Costory: cost control inside a delivery platform versus a standalone, flat-priced allocation layer.
- [Cast AI and Costory: optimization vs allocation](https://www.costory.io/compare/costory-vs-cast-ai): Cast AI automates Kubernetes optimization; Costory allocates and explains cost. They solve different problems and can run together.
- [CloudQuell vs Costory](https://www.costory.io/compare/costory-vs-cloudquell): CloudQuell and Costory compared on providers, allocation, assistants and pricing model.
- [CloudHealth vs Costory](https://www.costory.io/compare/costory-vs-cloudhealth): CloudHealth (Broadcom) vs Costory: governance-heavy FinOps suite vs an MCP-first FinOps harness. Who each is for.
### Alternatives
- [Alternatives to popular FinOps tools](https://www.costory.io/alternatives): Looking for an alternative to Vantage, CloudZero, Finout, Kubecost or AWS Cost Explorer? Options compared by fit, with claims checked October 2026.
- [Vantage alternatives in 2026](https://www.costory.io/compare/alternative-vantage): Options compared by use case (MCP access, pricing model, setup time), including when to stay on Vantage. Costory is one of the listed options and the publisher of the page.
- [CloudZero alternatives in 2026](https://www.costory.io/compare/alternative-cloudzero): Options compared by use case (MCP access, pricing model, setup time), including when to stay on CloudZero. Costory is one of the listed options and the publisher of the page.
- [Finout alternatives in 2026](https://www.costory.io/compare/alternative-finout): Options compared by use case (MCP access, pricing model, setup time), including when to stay on Finout. Costory is one of the listed options and the publisher of the page.
- [AWS Cost Explorer alternatives in 2026](https://www.costory.io/compare/alternative-aws-cost-explorer): Options compared by use case (MCP access, pricing model, setup time), including when to stay on AWS Cost Explorer. Costory is one of the listed options and the publisher of the page.
- [Cloudability alternatives in 2026](https://www.costory.io/compare/alternative-cloudability): Options compared by use case (MCP access, pricing model, setup time), including when to stay on Cloudability. Costory is one of the listed options and the publisher of the page.
## Use cases
- [All use cases](https://www.costory.io/use-cases): FinOps skills your agents run on Costory: allocation and showback, optimization, anomaly investigation, unit economics, budgets, finance close, AI spend. Costory finds and recommends the optimizations; your coding agent or your teams implement them.
### Allocation and showback
- [Allocate cloud costs without perfect tagging](https://www.costory.io/use-cases/cost-allocation-without-perfect-tagging): Map accounts, projects, resource names and partial tags to teams with rules, instead of waiting for a tagging project to finish.
- [One showback across AWS, Google Cloud, Azure, AI and data platforms](https://www.costory.io/use-cases/multi-cloud-showback): Teams see all their costs in one place, whatever the provider: cloud, Kubernetes, data platforms and LLM APIs.
- [Split shared infrastructure costs by actual usage](https://www.costory.io/use-cases/shared-cost-allocation): Allocate shared databases, networking and platform services to the teams that use them, based on real usage metrics.
- [Fill owner columns from a lookup file kept in git](https://www.costory.io/use-cases/cost-enrichment-tagging): Map namespaces, accounts or feature keys to teams with an enrichment file (beta), then report on the new columns in Explorer, dashboards, alerts and reports.
- [Split shared API costs across teams with usage data](https://www.costory.io/use-cases/shared-api-cost-split): Allocate untagged or shared API costs, like a single LLM key or a shared data platform account, to teams based on actual usage.
- [Split production and non-production costs automatically](https://www.costory.io/use-cases/automated-environment-allocation): Allocate every cost line to prod, staging, dev or sandbox using account names, projects, clusters and labels.
- [Tag coverage audit: the agent finds allocation gaps and drafts the rules](https://www.costory.io/use-cases/tag-coverage-audit): Find which label keys could become dimensions, how much spend each one covers, and where untagged cost is concentrated.
- [Kubernetes cost allocation by namespace, team and workload](https://www.costory.io/use-cases/kubernetes-cost-allocation): Break EKS, GKE and AKS costs down by cluster, namespace and workload, and roll them up to teams.
- [Dynamic cost reallocation for EKS using Datadog usage](https://www.costory.io/use-cases/eks-usage-based-reallocation): Split shared EKS cluster costs across teams in proportion to what each team's workloads actually use, measured in Datadog.
- [Stable team costs with contracted cost](https://www.costory.io/use-cases/contracted-vs-billed-cost): Give teams a usage-driven cost view that doesn't swing when Savings Plan discounts are shared across accounts.
- [Weekly showback per team, posted in each team's Slack channel](https://www.costory.io/use-cases/team-cost-reports): Every engineering team gets its own weekly cost report in its own Slack channel, scoped to the services it owns.
### AI and LLM spend
- [LLM cost allocation by team, feature and customer](https://www.costory.io/use-cases/llm-cost-allocation): Allocate model spend from OpenAI, Anthropic, Azure OpenAI, Bedrock, Vertex AI and OpenRouter to teams, features and customers.
- [All AI spend in one view: OpenAI, Anthropic, Bedrock, Vertex AI and more](https://www.costory.io/use-cases/ai-spend-overview): The agent tracks total AI spend across model APIs, cloud AI services and AI tools, and reports the weekly trend in one place.
- [Cost per AI feature: tokens, GPUs and infrastructure in one number](https://www.costory.io/use-cases/cost-per-ai-feature): Combine model API spend with the GPU and serving infrastructure behind each AI feature, then divide by usage for a true cost per request.
- [Choose LLM models and routing based on cost per feature](https://www.costory.io/use-cases/llm-model-routing): The agent tracks per-model spend for each AI feature across OpenAI, Anthropic, Bedrock and Vertex AI.
- [Cursor and Claude Code spend per developer](https://www.costory.io/use-cases/ai-coding-spend-per-developer): Track AI coding tool spend per user and per model, from Cursor and the Anthropic Claude Code connector, next to the rest of your cloud.
- [claude.ai Enterprise cost per user](https://www.costory.io/use-cases/claude-enterprise-cost-per-user): Import per-user Claude Enterprise cost with an Analytics API key, and allocate seats and usage to teams.
- [Track generative media spend: ElevenLabs, Kling AI and BytePlus](https://www.costory.io/use-cases/genai-media-costs): Import voice, video and media generation spend next to cloud and LLM costs, and allocate it to products.
### Anomalies and investigation
- [Investigate a cost spike and find the deploy behind it](https://www.costory.io/use-cases/cost-spike-investigation): When spend moves, the agent finds the deploy, team and service behind it, checks incidents and provider events from the same window.
- [Investigate anomalies in a Slack thread with Costory](https://www.costory.io/use-cases/slack-anomaly-threads): When an alert fires in #finops or a team channel, reply in the thread and Costory investigates: which team, which service, which deploy.
- [Cost impact of every deploy, from GitHub events](https://www.costory.io/use-cases/deploy-cost-impact): Capture deploys, PR merges and releases from GitHub or your CI/CD pipeline, and see which ones changed cost.
- [Add the cost of an incident to your postmortem](https://www.costory.io/use-cases/incident-cost-postmortem): Incidents are logged as events. After each one, the ag