Mint a single-use upload URL for pushing a conversation-attached PDF to Bankstatemently before converting it. Use this ONLY when you have no other way to reference the attached file (no pdf_file/pdf_url equivalent for this host) — e.g. a code-execution sandbox that can see the file on disk but has no URL for it. Playbook: (1) check your sandbox's uploads/attachments directory first — if the file isn't there yet, the mount can lag behind the conversation; ask the user to re-attach or wait a moment and check again before calling this tool. (2) Call request_upload to get upload_url and upload_id. (3) PUT the raw PDF bytes to upload_url with header Content-Type: application/pdf, e.g.: curl -X PUT "<upload_url>" -H "Content-Type: application/pdf" --data-binary @<path-to-file>. (4) Once the PUT succeeds, call convert_statement with upload_id set to the same value — never pdf/pdf_url/pdf_file for this flow. The URL and token are single-use and expire quickly; call request_upload again for a fresh one if the PUT fails partway through — never retry a failed PUT against the same URL. If the PUT fails with a network error or a "host not allowed"-style denial, the sandbox is likely blocking outbound requests to api.bankstatemently.com — tell the user to add api.bankstatemently.com to their host's code-execution allowed-domains setting (on claude.ai: Settings → Capabilities → Code execution) and retry. To convert several statements at once, pass count (1-100) instead of calling this tool once per file: the response returns "uploads", an array of that many { upload_id, upload_url } pairs — PUT each file to its own upload_url, then make ONE convert_statement call with upload_ids set to every upload_id. Free to use — no credits consumed (conversion itself still costs credits, same as any other convert_statement call).

Convert a bank statement PDF into structured data or a spreadsheet. When the user attaches a PDF in the conversation, it arrives automatically as pdf_file — never encode it yourself. Otherwise, pass pdf_url for a public HTTPS link. If your host has no way to reference the attached file at all (no pdf_file/pdf_url equivalent), call request_upload first and pass its upload_id here instead. The base64 pdf parameter is a last resort only, for a caller with no other way to reference the file. To convert several statements in one call, pass upload_ids (the array from a single request_upload call made with count set) instead of pdf/pdf_url/pdf_file/upload_id — mutually exclusive with those four. This batch form only ADMITS each file (queues it, or reports an already-completed duplicate) and returns immediately with a compact per-file status list plus a summary — it never waits for conversion, so call get_statement per document_id once ready rather than expecting inline results here. Returns accounts, transactions, and metadata. output_format "json" (default) returns the data inline, renderable in chat. The other formats (csv, xlsx, qbo, xero) return a time-limited download link instead: present it as a normal link. Every response includes a "summary" field: use it as the single source of truth for what happened. If the conversation is not in English, translate it faithfully into the conversation language; never add details it doesn't contain. Never echo raw status values (e.g. "completed") or field names. Consumes credits (1 per page). Page limit depends on your plan.

Fetch the full converted data for a previously processed document. Use this after convert_statement returns a "processing" status, or to re-fetch results. output_format "json" (default) returns the data inline, renderable in chat. The other formats (csv, xlsx, qbo, xero) return a time-limited download link instead: present it as a normal link. data_mode selects which projection of the data you get: omit it for each output_format's existing default behavior. "normalized" is the cleaned, interpreted view; "original" includes each transaction's raw column values exactly as printed on the source PDF (originalData); "enhanced" is a reformatted view of the original columns (csv/xlsx only for now). Fetch data_mode: "original" when you plan to submit results to evaluate_benchmark — pass its originalData through verbatim; an absent originalData scores that benchmark's raw-fidelity dimension 0 for this document. Every response includes a "summary" field: use it as the single source of truth for what happened. If the conversation is not in English, translate it faithfully into the conversation language; never add details it doesn't contain. Never echo raw status values (e.g. "completed") or field names.

Run AI transaction categorization on a previously processed document, then return its category mappings. Returns cached categories with no charge if this document was already categorized. Consumes credits (pooled per page, same rate as the categorize toggle on the website) the first time — free on every re-fetch after. Every response includes a "summary" field: use it as the single source of truth for what happened.

Browse your previously converted bank statements with pagination and optional status filter.

Hide a failed, rejected, or cancelled document from future list_statements results. Use this only when the user asks to clear a terminal failed/rejected/cancelled conversion from their history. This is not a delete: it marks the document dismissed and leaves stored data/artifacts untouched.

Your remaining Bankstatemently credits — the processing quota, NOT credit/debit transactions. Use for: how many credits do I have, remaining pages, plan limits, quota, how many pages can I upload. 1 credit = 1 page of bank statement processing. Also reports your plan's operational limits (max pages per upload, max upload size, daily spend cap) so you can size a multi-file batch correctly before starting it.

Report how well a previously converted bank statement was parsed: submit a 1-5 rating, optionally with structured feedback (only accepted when the rating is 3 or below) and use-case tags. Calling this again for the same document updates your existing rating without clearing feedback already submitted for it. Returns the stored rating state in the response — there is no separate tool to read your own rating back. Every response includes a "summary" field: use it as the single source of truth for what happened.

Score parsed bank statement transactions against the Bankstatemently benchmark ground truth. Accepts a statement_id (e.g. "bsb-001") or content_hash, plus your parsed transactions. Returns extraction accuracy, integrity score, and an overall score. Only statements marked published: true in the catalog can be evaluated — held-out statements return an error. transactions[].originalData is optional but strongly recommended: fetch it via get_statement with data_mode: "original" and pass it through verbatim — an absent originalData scores that transaction's raw-fidelity (parsed) dimension 0; never fabricate a value. Free to use — no credits consumed. Read the benchmark://catalog resource first to see available statements and their published status.

A transaction is a single line as printed on one account's statement — one side of any movement. Return a filtered list of transactions across your converted statements, capped at 50 rows. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range. Every response names the scope it actually evaluated (document count + covered date range) and each returned row carries its source document's content_hash so you can cite it. For "how many credits do I have" / processing quota / remaining pages, use get_credits instead — that is not a transaction.

Compute a single metric (sum/average/count/max/min) over a filtered set of transactions across your converted statements. Results are per-currency — never sum across currencies yourself. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range. For "how many credits do I have" / processing quota / remaining pages, use get_credits instead — that is not a transaction.

Group transactions by a dimension (month/category/merchant/account/currency) and apply a metric to each group. Results are per-currency. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range.

Return the top N groups ranked by metric (descending), per-currency for monetary metrics. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range.

Side-by-side metric comparison for two filtered groups of transactions (e.g. one category vs another, one month vs another). Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range.

Compute a time series by grouping transactions into week or month buckets and applying a metric — useful for trends. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range.

Match transfers between your own accounts. A transfer is TWO transactions — a debit leaving one of your accounts and a credit arriving in another — matched as two sides of the same movement (amount and date aligned); account-level successions (an account closing into a successor) are matched too. A payment to an outside party is not a transfer here: only movements with both sides visible in your statements are matched. THE way to answer any "was money moved between my accounts" / "did I transfer X" question — never try to answer a money-moved-between-accounts question with list_transactions + arithmetic; always call this tool instead. Scope defaults to all your completed statements; pass "scope" to narrow to specific accounts/products and/or a date range. Every response reports the match window (in days) it used, even when no transfers are found — a lack of matches is never silent about how hard it looked. To find large movements with NO matching counterpart in your other accounts — e.g. "trace transfers over $10,000; which ones leave without a known destination?" — pass "amountMin": reconciled pairs and successions are filtered to that floor, and the response gains an "unmatched" bucket of large movements (debits leaving, or unexplained credits arriving) with no matching pair, candidate, or succession. Omit amountMin for the ordinary reconciled-pairs answer.