Turn bank statement PDFs, CSVs, XLSX and OFX into categorised transactions plus a summary.
Scored across 3 tools
Each tool maps to a distinct stage of the pipeline: starting a conversion, polling its async progress, and retrieving the final categorized transactions. There is no meaningful overlap between them.
All tool names follow a clear lowercase verb_noun snake_case pattern: convert_statement, get_job, get_transactions. The naming clearly signals each tool's action and target.
Three tools is well-scoped for the narrow Statement to Budget domain. Each tool is necessary for the async conversion workflow and none feel redundant or missing as a primary action.
The core lifecycle—submit, poll, fetch results—is covered completely. The only minor gap is that the preview/unlock flow lacks an explicit refresh or re-fetch tool to reflect full access after a human completes the purchase.
Upload a bank statement (PDF, CSV, XLS, XLSX, or OFX) as base64 and run AI extraction + categorisation. Returns a session_id plus access.level: "full" when an API key with credit was supplied, otherwise "preview" (5 sample rows, other rows masked) with an unlock_url a human can open to buy the full file for $9.99. Large statements can take 1-3 minutes; if status is "processing", poll get_job.
| Name | Required | Description | Default |
|---|---|---|---|
| filename | Yes | Original file name including extension | |
| password | No | PDF password if the file is locked | |
| content_base64 | Yes | File bytes encoded as base64 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full disclosure burden and meets it: it reveals the two access levels ('full' vs 'preview'), the 5-sample-row limitation, the $9.99 unlock_url, the 1-3 minute latency for large statements, and the 'processing' status path. This gives an agent a realistic model of the tool's behavior beyond the schema.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is compact and information-dense: three sentences cover input format, return contract, pricing, access-level semantics, latency, and polling behavior. Nothing is wasted, and the primary action is front-loaded in the first sentence.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with no output schema and no annotations, this description is remarkably complete: it specifies accepted inputs, required encoding, response contents, access-level behavior, pricing, expected duration, and the async status path. An agent has enough context to invoke it correctly and handle the outcome.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so the baseline is 3, but the description adds meaningful value by enumerating accepted file extensions (PDF, CSV, XLS, XLSX, OFX) and reinforcing that content is sent as base64. This supplements the filename and content_base64 parameter meanings beyond the schema's brief descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description names a specific action (upload and run AI extraction + categorisation) on a specific resource (a bank statement file) and states its outputs (session_id, access.level). It also distinguishes itself from siblings by positioning itself as the upload/conversion entry point and explicitly telling the agent to poll get_job when status is 'processing'.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives clear context for when to use this tool: uploading a statement in accepted formats. It names the sibling get_job as the polling follow-up when processing is incomplete, but it does not explicitly explain when to use get_transactions instead. This is clear enough for most cases but stops short of explicit when-to-use versus all alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Poll a conversion job by session_id. Status is processing, complete, or error.
| Name | Required | Description | Default |
|---|---|---|---|
| session_id | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must carry the behavioral burden. It does disclose that the status can be processing, complete, or error, which is useful. However, it does not describe the response shape, error details, or whether repeated polling is expected beyond the implication of the word 'Poll'.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two short sentences with no filler, front-loading the action and the key statuses. Every word contributes to understanding the tool's core behavior.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple one-parameter polling tool, the description is close to sufficient, and listing the statuses hints at the return value. However, without an output schema or annotations, it should also mention where session_id comes from and what a complete or error response contains.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description must compensate. It does indicate that session_id identifies the conversion job, adding a bit of meaning beyond the bare string type. It does not explain how to obtain session_id, such as from a convert_statement response.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states a specific verb ('Poll') and a specific resource ('a conversion job'), and it lists the possible statuses, so the tool's purpose is clear. It does not explicitly differentiate from sibling tools like convert_statement or get_transactions, but the function is distinct enough by name and phrasing.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives no guidance on when to use this tool versus convert_statement or get_transactions, and it does not mention prerequisites such as where session_id comes from. The word 'Poll' only implies that this is for checking job status after submission, but the tool does not say that explicitly.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Fetch categorised transactions and a spending summary for a completed session. Preview sessions return 5 sample rows, locked_rows (date, category, masked description) for the rest, and unlock_url.
| Name | Required | Description | Default |
|---|---|---|---|
| session_id | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations available, the description carries the full transparency burden and does a solid job: it discloses conditional behavior for preview sessions versus completed sessions, mentions the 5-sample-row limit, locked_rows, and unlock_url. It does not cover error cases or authentication, but for a simple fetch tool this is meaningful behavioral detail.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two compact sentences: the first front-loads the core function, and the second adds the key behavioral nuance. Every phrase adds information, and there is no redundant or filler wording.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a one-parameter tool with no output schema, the description covers the main call contexts, including completed sessions and preview sessions, and lists the returned components. It could be more complete by specifying expected error behavior for invalid session IDs, but the essential usage context is present.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has zero description coverage, and the tool description does not explain the format, origin, or constraints of session_id beyond implying it identifies a session. The session concept is indirectly clarified by the completed/preview distinction, but that is too weak to compensate for a completely undocumented required parameter.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states a specific verb ('Fetch') and resource ('categorised transactions and a spending summary for a completed session'), which makes the tool's purpose immediately clear. It also distinguishes from siblings like get_job by focusing on transaction data and session completion state, including concrete preview behavior.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives clear context that the tool is intended for completed sessions and separately describes what happens for preview sessions. It does not explicitly compare against convert_statement or get_job, but the session-specific wording is enough to guide typical use.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
convert_statementget_jobget_transactionsTurn bank statement PDFs into categorized, balance-checked transactions and reports.
Bank statement PDFs become CSV or JSON transactions with dates, amounts and balances.
Convert PDF bank statements into structured transactions, accounts, and balances.
Convert PDF bank statements to checked Excel, CSV or JSON with balance validation.
Add one secure layer between your agents and this server.
