Auto-build interactive bank expense reports with n8n + OpenAI

Upload CSV or text PDF → parse transactions → AI-categorize merchants → generate HTML report with charts, monthly breakdowns, and recurring-charge detection

Who

Individuals or small teams with an n8n instance who need automated bank statement analysis

You get

An interactive web report with category bars, monthly tables, annualized recurring costs, AI commentary, and CSV download

You need

n8n

Steps

1. Configure trigger and upload form

   After importing the workflow into n8n, locate the 'Upload Statement' form trigger node. It is pre-configured to accept .csv, .txt, .pdf files. No modification is required, but you may adjust the form title or description. Activate the workflow, then copy the production URL from this node for later uploads.

2. Detect file type and extract text

   The workflow uses the 'Is it a PDF?' IF node to check the uploaded file's MIME type. If PDF, it routes to 'Extract PDF Text'; otherwise to 'Extract CSV Text'. Both branches merge at the 'Combine File Branches' node, outputting a single raw text string.

3. Parse transactions

   The 'Parse Transactions' code node auto-detects delimiters (comma/semicolon/tab/pipe), date format (DD/MM or MM/DD), and column layout, outputting a normalized transaction list and deduplicated merchant list. Supports US, UK, Indian, and European formats. Throws explicit errors if no transactions can be read.

4. AI-categorize merchants

   The 'Build Category Prompt' code node constructs a prompt sending the deduplicated merchant list to the OpenAI model. The model assigns each merchant a fixed category (e.g., Housing, Groceries, Subscriptions) and a readable name. The 'Categorize Payees' AI Agent node then executes this, instructing the model to return pure JSON only, never invent merchants, and never perform arithmetic.

   Execution prompt · Source text。Categorize Payees node systemMessage field

   Copy prompt

   You categorise bank statement payees. You are given messy abbreviated merchant strings and you assign each one a category from a fixed list, plus a readable name.
   
   You never perform arithmetic, never invent merchants, and never return a merchant that was not in the input list.
   
   You always respond with a single valid JSON object and nothing else. No markdown code fences, no preamble, no explanation.

   Execution prompt · Editorial suggestion。Build Category Prompt node jsCode-constructed user prompt; {{merchant_list}} is runtime variable

   Copy prompt

   Below is a list of merchant names taken from a bank statement. Assign each one a spending category.
   
   Allowed categories, use these exact strings:
   Housing, Groceries, Dining & Takeaway, Transport, Fuel, Utilities, Phone & Internet, Subscriptions, Shopping, Health & Medical, Insurance, Education, Entertainment, Travel, Fees & Charges, Cash Withdrawal, Transfers, Income, Business Expense, Other
   
   Guidance:
   - Income is only for money coming in such as salary, client payment, refund, interest.
   - Subscriptions means recurring digital or membership services such as streaming, software, gym.
   - Transfers means moving money between own accounts, or person to person transfers with no clear purpose.
   - Fees & Charges means bank fees, interest charged, penalties, late fees.
   - Use Other only when the name genuinely gives no clue. Do not overuse it.
   - The merchant names are messy and abbreviated. Infer from partial names where a reasonable person could.
   
   Merchants:
   {{merchant_list}}
   
   Return strictly this JSON object and nothing else:
   {
     "categories": [
       { "merchant": "the exact merchant string given above", "category": "one of the allowed categories", "clean_name": "a readable brand or payee name, title case, or the merchant string cleaned up if unknown" }
     ]
   }
   
   Rules:
   - Return one entry for every merchant listed, using the exact merchant string as given.
   - Never invent merchants that were not listed.
   - Do not use em dashes anywhere in the output.
   - No markdown code fences, no explanation before or after the JSON.

5. Calculate summary data

   The 'Calculate Totals' code node computes: spend/income totals, category percentages, monthly breakdowns, largest single transactions, recurring charge detection (based on merchant, amount stability, and similar charge dates), and annualized costs. Uncategorized transactions default to Other or Income.

6. Generate AI commentary

   The 'Build Insights Prompt' code node formats the computed summary into a prompt for the 'Insights Agent' AI Agent node. The model generates a plain-language summary, observations (tagged Watch/Notable/Positive), a recurring-charge note, and questions to verify. The model is instructed not to give investment advice, not to invent numbers, and not to assume trends.

   Execution prompt · Source text。Insights Agent node systemMessage field

   Copy prompt

   You read a pre-calculated bank statement summary and describe what it shows, in plain language.
   
   Every figure you are given was computed arithmetically. You treat those figures as exact, quote them as given, and never recalculate or contradict them. You never invent a number, merchant or trend that is not in the data.
   
   You describe what the data shows. You do not give investment advice, do not recommend financial products, and do not tell the person what they should or should not buy.
   
   You always respond with a single valid JSON object and nothing else. No markdown code fences, no preamble, no explanation.

   Execution prompt · Editorial suggestion。Build Insights Prompt node jsCode-constructed user prompt; {{variables}} are runtime substitutions

   Copy prompt

   Here is a summary of one bank statement. Every figure below was calculated arithmetically from the statement, so treat all numbers as exact and never recalculate or contradict them.
   
   Period: {{period_start}} to {{period_end}} ({{month_count}} month(s), {{transaction_count}} transactions)
   Total spent: {{spend_total}}
   Total received: {{income_total}}
   Net: {{net_total}}
   Average monthly spend: {{avg_monthly_spend}}
   
   ## Spend by category
   {{category_lines}}
   
   ## Month by month
   {{month_lines}}
   
   ## Repeating charges detected (same payee, near identical amount, similar day each month)
   {{recurring_lines}}
   
   ## Repeating charges that stopped before the final month
   {{dormant_lines}}
   
   ## Largest single transactions
   {{biggest_lines}}
   
   Write a short, useful read of this statement. Return strictly this JSON object:
   {
     "summary": "3 to 4 sentences describing where the money actually went this period, in plain language, quoting the real figures",
     "observations": [ { "title": "short headline", "detail": "2 to 3 sentences citing the exact figures from above", "tag": "Watch|Notable|Positive" } ],
     "subscription_note": "2 to 3 sentences about the repeating charges, including the total annualised cost, and specifically calling out any that stopped mid period since those may be forgotten or cancelled. If none were detected, say so plainly.",
     "questions_to_check": [ "a specific thing the account holder should verify on their own statement, phrased as a question" ]
   }
   
   Rules:
   - Give 4 to 6 observations and 3 to 5 questions_to_check.
   - Only use the figures given above. Never invent a number, a merchant, or a trend you cannot see in this data.
   - If the period is a single month, do not describe trends over time.
   - Describe what the data shows. Do not give investment advice, do not recommend financial products, and do not tell the person what they should or should not buy.
   - Categories were assigned automatically from payee names and can be wrong, so do not treat any single categorisation as certain.
   - Do not use em dashes anywhere in the output.
   - No markdown code fences, no explanation before or after the JSON.

7. Build and serve HTML report

   The 'Build Report Page' code node assembles an interactive HTML page: summary tiles, category bars, monthly table (when multi-month), recurring charges table (with 'stopped' badges), largest transactions table, AI commentary section, full transaction filterable table, and CSV download link. The 'Show Report' form node returns this HTML directly as the response.

You are an n8n workflow setup assistant. The user wants to replicate the 'Bank Statement to Expense Breakdown' template. Assist in this order:
1. Confirm the user has an n8n instance (self-hosted or cloud) and an OpenAI API key;
2. Guide them to import the workflow JSON from https://n8n.io/workflows/17753/;
3. Configure the API credential in the 'OpenAI Chat Model' node, ensuring both AI Agent nodes connect to it;
4. Activate the workflow, copy the 'Upload Statement' form URL, and test with a statement;
5. Remind them to use only CSV or text-based PDFs—scanned documents are not supported;
6. If customizing categories, guide editing the category list inside the 'Build Category Prompt' node.