nkfix-mcp
Related Servers
Alternatives to nkfix-mcp
No user-submitted related servers found.
Related Servers
- AlicenseAqualityBmaintenanceEnables AI assistants to read and work with sevDesk invoices, receipts, bank transactions, and reports through tiered permissions, with human confirmation for legally binding actions.311MIT
- FlicenseAqualityDmaintenanceConnects AI assistants to the Lanten property management API, allowing them to manage tenants, units, and work orders using natural language.1511 npm-
- AlicenseBqualityBmaintenanceEnables AI assistants to query and manage self-hosted accounting data—invoices, balances, and books—through natural language, with read-only tools by default and optional scoped write operations.3024 PyPI1AGPL 3.0
- FlicenseNot gradedqualityBmaintenanceConnects AI assistants to the official DATEV APIs so users can query bookkeeping, payroll, and documents in plain language; any write is previewed and requires a matching one-time confirmation before it is executed.-
- AlicenseAqualityBmaintenanceEnables AI assistants to interact with the nula.bg accounting platform — invoices, purchases with OCR, customers, inventory, and bank feeds — through an API key, with read-only mode by default.1335 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to read and draft accounting documents such as invoices, vouchers, contacts, and articles via the Lexware Office API, including PDF rendering and receipt uploads. Supports OAuth and static token authentication for secure access.MIT
TDQS
Scored across 25 tools
Most tools target a distinct resource (properties, apartments, tenants, invoices, meters, statements) or a clearly differentiated action (check vs record vs upload). The diagnostic tools overlap somewhat: get_data_quality, get_billing_year_status and precheck_statement all report missing readings/invoices and deadlines for a property, so an agent may hesitate between them. Descriptions provide enough cues (precheck is tenant+period bound, year status is per property-year) to keep them separable.
Every tool is snake_case with a clear verb_noun pattern: list_*, get_*, check_*, record_*, upload_*, precheck_statement. The verbs are used consistently by intent (read vs validate vs write), and there are no mixed camelCase or vague standalone names.
25 tools is at the heavy end for a single billing domain, and the read surface is dense: 14 list_* plus three overlapping diagnostic endpoints. The domain is genuinely broad (properties, tenants, costs, meters, statements, deadlines), but a few tools could be merged without losing capability.
The write surface is deliberately append-only (record_payment, record_meter_reading, record_invoice, upload_invoice_document) and read coverage is strong, but there is a notable dead end: statements can be listed, inspected and prechecked but never created, and no tool exists to add tenants, properties, cost categories or to correct/update existing records. Many workflows therefore require manual human action mid-stream.