How to Download AWS Invoices (Single Account and Organizations)

AWS has been around for long enough that its billing surface has been redesigned three times. In the current version, invoices live under Billing and Cost Management, which was itself renamed from the old Billing console, which replaced an even older Billing dashboard. If you have been on AWS since 2015, you have clicked "where is the bill" through all three incarnations. The data underneath is the same. The path to the PDF keeps moving.

There are three documents that matter for AWS bookkeeping. The monthly invoice PDF is the high-level bill, suitable for your accounting software. The Cost and Usage Report (CUR) is the granular CSV or Parquet file, the one you want for FinOps, chargeback, or resource-level analysis. The credit memo is the document AWS issues when a credit or refund is applied, which is a separate PDF from the regular invoice and frequently missed. For most companies, the monthly invoice is enough. For anyone with real cloud spend (over a few thousand dollars a month), CUR is essential.

This guide walks through how to get each document, how the flow differs for single accounts versus AWS Organizations with consolidated billing, and what automation looks like for teams that run more than one AWS account.

The manual way: downloading AWS invoices from the console

Three different paths depending on which document you need.

Step 1: find the right billing console

Log in to the AWS Management Console. In the top-right account menu, click Billing and Cost Management. You land on the Billing Dashboard.

As of 2026, the left sidebar has a Payments section with Invoices as a sub-link. Older accounts may see a top-level Bills link with a different layout. Both get you to the same data. If you are logged in to a member account inside AWS Organizations, you will not see Invoices at all. Member accounts have no billing of their own; everything rolls up to the management (payer) account. Switch to the management account if you are trying to retrieve the invoice.

Step 2: download the monthly PDF

On the Invoices page, select the billing period. Each month's invoice shows up as a row with the date, total charged, and a Download button. Click it to get the PDF.

AWS generates the monthly invoice on the 2nd or 3rd day of the following month. Before that date, the current month's charges show up as "Bill (in progress)" with no downloadable PDF yet. Do not try to pull January's invoice on February 1st; it will not be available until the 2nd or 3rd.

The PDF includes the account number, billing period, a breakdown by service group (compute, storage, database, networking, and so on), applied credits, taxes, and the total due. It is suitable for accounting software imports and for paying the bill if you are on invoice-based payments rather than auto-charged cards.

Step 3 (Organizations): grab the consolidated invoice

If you use AWS Organizations with consolidated billing (the default for multi-account setups), the management account's monthly invoice lists every member account's charges as individual line items. The total is the sum across all accounts.

Download it the same way: Billing and Cost Management → Payments → Invoices → click the month's row → Download. The PDF can be 10 to 100+ pages for larger organizations, because each member account is broken out separately.

If you need per-account invoices for internal chargeback, AWS does not generate those automatically. You either need to build your own chargeback reports from the Cost and Usage Report (step 4) or use AWS's built-in Cost Categories feature to split spend by business unit and report on those categories separately.

Step 4: set up the Cost and Usage Report (CUR)

The CUR is a separate setup, not a one-click download. Go to Billing and Cost Management → Cost and Usage Reports → Create report. Configure:

• Time granularity: hourly, daily, or monthly. Most FinOps workflows use hourly for compute cost analysis.
• Include resource IDs: enables per-resource cost tracking (e.g., which specific EC2 instance cost what).
• Data refresh settings: whether to refresh the CUR when older data changes (refunds, credits).
• S3 bucket destination: where AWS drops the CUR files. You need to create the bucket first and apply the right IAM policy.

Once configured, AWS writes the CUR to S3 up to three times a day. The files are large: a mid-sized account with 30 days of hourly data generates a few GB of Parquet files per month. Pull them into a data warehouse or Athena to query.

CUR is the data source for AWS QuickSight Cost Intelligence Dashboards, for third-party FinOps tools like Vantage or CloudHealth, and for any custom chargeback logic you build.

Step 5: credit memos and refunds

When AWS issues a credit (enterprise support credits, migration credits, outage reimbursements, billing error refunds), the credit shows up on the next monthly invoice as an "Applied credits" line. For larger credits (over a few hundred dollars), AWS also emails a separate credit memo PDF to the billing contact. This memo is a distinct document from the invoice and should be filed alongside it.

If you have received credits and they are not reflected on your invoices, check Billing and Cost Management → Credits. You can see every credit issued, its expiration date, and how much has been applied to date.

Step 6: rename and file

AWS's PDF invoice is named Invoice-{accountNumber}-{month}-{year}.pdf, which is unambiguous but not human-readable. Rename with date, vendor, and total: 2026-04-01_aws_$12580.pdf. For Organizations, include the management account's company name: 2026-04-01_aws_acme-mgmt_$45200.pdf.

For CUR files, the S3 path naming is already structured by AWS (by date, report name, manifest). Leave those as-is and query them via Athena or a data warehouse rather than trying to file them like PDFs.

Why manual breaks at scale

For a single AWS account spending under a few thousand dollars a month, manual invoice downloads are fine. Pull the PDF once a month, drop it in a folder, done.

For everyone else, the problems compound quickly.

Multi-account tracking without consolidated billing. Some organizations run separate AWS accounts for different environments (production, staging, development) or different business units, without enabling Organizations consolidated billing. That means one invoice per account per month. A mid-sized company with 10 AWS accounts generates 10 invoices a month, each needing its own download. If any of those accounts are on different payment methods, reconciling against bank feeds becomes a multi-hour task.

Credit memo tracking. AWS credits apply automatically to the invoice, but the credit memo PDFs arrive as separate emails, sometimes days apart from the invoice itself. If you archive only the monthly invoice and not the credit memo, your audit trail is incomplete. In an audit, the question "what was this credit for?" requires the memo, and finding it months later is painful.

Cost Explorer doesn't replace invoices. Cost Explorer is a great analysis tool, but it is not a tax document. Some finance teams mistakenly screenshot Cost Explorer and attach it to expense reports. This works until an auditor asks for the official invoice and you have to backfill six months of downloads.

Rapid month-over-month cost changes. Cloud spend is volatile. A service that cost $500 last month might cost $5,000 this month because someone left a large instance running. The invoice shows the total, but the root cause requires CUR analysis. Without CUR data, your finance team can tell you the bill went up but not why. This is the difference between knowing you spent too much and knowing what to fix.

Manual AWS bookkeeping is one of those tasks that starts small and grows with the business. The five-minute monthly task becomes an hour as the organization adds accounts, then a multi-hour task once CUR analysis enters the picture, then a full-time-equivalent role once FinOps becomes a function.

Manual vs automated

Automating with Inbox Ledger

AWS is a good candidate for hybrid automation. The monthly invoice PDF works well through an inbox-capture flow. The Cost and Usage Report is better handled separately (AWS delivers it to S3; your data team consumes it from there). Inbox Ledger focuses on the former.

Connect the inbox where AWS sends billing notifications. By default AWS emails the account's billing contact. For multi-account setups, either the management account's billing contact gets notifications for every account, or each account has its own billing contact. Connect the inbox via Gmail OAuth, Outlook OAuth, or IMAP. Inbox Ledger pulls historical AWS emails (default 90 days) and runs incrementally.

Capture the notification email and the linked PDF. AWS's billing notification email says "Your AWS account bill is available" and links back to the console. The extractor recognizes this email type, uses the ticket number or invoice reference to identify which invoice it represents, and (via integration with AWS's Billing API, or via a manual one-time download to the archive) attaches the PDF to the record. The structured data (account number, billing period, total) comes out of the email body or the PDF.

Match credit memos to invoices. Credit memo emails from AWS reference the original invoice by number and period. The extractor links them automatically, so your archive shows the invoice with its associated credit as a single logical record. No more hunting for "what was this $500 credit for?" a month later.

Handle multiple accounts cleanly. If the billing inbox receives notifications for 10 different AWS accounts, each gets tagged with its account number at extraction. Filter the archive by AWS account to see only that account's history, or view the whole org's history together.

Push to accounting software or a FinOps tool. Every invoice can post as a bill in QuickBooks or Xero, with the appropriate GL mapping (usually "Cloud Services" or "Infrastructure"). For teams using Google Sheets for monthly IT spend tracking, rows can append automatically. The Cost and Usage Report is not in scope for email-based automation; connect that directly from S3 to your data warehouse.

For an AWS-specific setup walkthrough, including how to configure the billing contact, how to enable consolidated billing for Organizations, and how to set up the Cost and Usage Report with the right IAM permissions, see the AWS portal page.

The integrations page has the full list of AP destinations including QuickBooks and Xero. The AI processing page covers how the extractor handles AWS-specific edge cases like partial-month prorations (when you enable or disable a service mid-month), Reserved Instance amortization, and Savings Plans credits. If you spend a lot of time in the AWS console, the Chrome extension has a one-click "Send to Inbox Ledger" button that works on the Invoices page. For teams that consolidate reporting in spreadsheets, the Google Sheets integration tool page has the setup walkthrough for monthly rollups.

Gotchas and edge cases

AWS has a lot of them. Here are the ones that burn finance teams most often.

Reserved Instances and Savings Plans amortization. When you buy a Reserved Instance or Savings Plan upfront, AWS charges you the full amount on that month's invoice. For accounting purposes, most teams want to amortize the cost over the one- or three-year term, not book it all in the month of purchase. The invoice shows the full payment; your accounting entries need to split it. CUR data is the usual source for amortized cost per month.

Cross-region charges and data transfer fees. Data transferred between AWS regions, or from AWS to the internet, appears on the invoice under "Data Transfer" as a single line, even though it might represent dozens of services. Finding out which service is the main contributor requires Cost Explorer or CUR analysis. The invoice alone will not tell you.

AWS Marketplace charges. Third-party software bought through the AWS Marketplace (Datadog, MongoDB Atlas, New Relic, etc.) appears on your AWS invoice as a Marketplace line item. For accounting purposes these are usually categorized as software subscriptions, not AWS infrastructure. Some finance teams want these broken out separately. The Marketplace section of the invoice lists each vendor; the extractor can tag these with a separate category for downstream routing.

Tax jurisdiction changes. AWS collects sales tax or VAT based on the account's billing address. If you move your company or add a new business entity, the tax lines on your invoice will change accordingly. For international operations, this can mean a single month where you are charged under two tax regimes if the change happened mid-month.

Enterprise Discount Program credits. If you have an Enterprise Discount Program (EDP) commit with AWS, your monthly invoice shows list prices before the commit discount is applied, and then a separate line item for the EDP discount at the bottom. Your accounting entries should use the net-of-discount figure, not the list prices. Make sure your extraction tool captures the discount line and calculates the net.

Consolidated billing rounding. When AWS Organizations consolidates billing across many member accounts, rounding differences between individual account calculations and the consolidated total are usually under a dollar but occasionally larger. Do not spend an hour reconciling a $0.23 difference; check that it is within AWS's documented rounding tolerance.

Invoice delivery delays in regulated regions. AWS accounts in certain regulated regions (China via AWS China, GovCloud, etc.) have different invoice generation schedules. The China accounts are operated by a separate entity (Sinnet or NWCD) and their invoices are not in the main AWS billing console. If your company has multi-region AWS presence, make sure you are checking the right billing console for each.

When automation is not worth it

Solo developers or small startups with a single AWS account spending under $500 a month do not need to automate. The monthly PDF download is a two-minute job. Spend the time on product, not FinOps tooling.

Companies using AWS solely through a managed service provider (MSP) also do not need direct invoice automation. The MSP bills you; AWS bills them. Your accounting interacts with the MSP's invoices, not AWS's raw invoices. Make sure the MSP is passing through the granular cost data if you need it for chargeback.

Automation is worth it when you have multiple AWS accounts (consolidated or not), when AWS spend is over a few thousand dollars a month and needs real analysis, or when you have a finance team that needs clean invoice data in QuickBooks or Xero without manual PDF handling. At that scale, the monthly manual process is a real time cost and the risk of missing a credit memo or a weird tax line is non-trivial.

For FinOps-specific analysis (per-team chargeback, RI utilization, rightsizing recommendations), invoice automation alone is not enough. You need CUR data in a data warehouse and a proper FinOps tool on top. Inbox Ledger handles the accounting side; your data platform handles the optimization side.

Closing: separate the accounting bill from the FinOps data

AWS's billing surface has two legitimate uses: paying the bill each month and understanding what to optimize next. The monthly PDF invoice serves the first purpose. The Cost and Usage Report serves the second. They are different documents for different workflows and should be automated differently.

For the accounting side, connect the inbox that receives AWS billing notifications, let the extractor pull invoice data and credit memos into a unified archive, and route the structured output to QuickBooks, Xero, or whatever your AP tool is. Start on the integrations page, pick a destination, and let a month of AWS invoices flow through. The monthly PDF handling will drop off your month-end checklist.

For FinOps, configure the CUR to land in S3 and wire it to your data warehouse. That is a separate workflow with different tooling, and it is worth investing in if your AWS spend has crossed into the "this matters financially" range.