Model Context Protocol server for Lodgify Public API - Expose Lodgify endpoints as MCP tools for AI assistants
The Lodgify MCP server presents mixed quality across 9 tools. Strengths: all tools have descriptions (10 - 450 chars, well-formed); parameter schemas are present with Zod validation; input sanitization and error wrapping are implemented. Weaknesses: descriptions lack depth for several tools (lodgify_get_quote has NO visible description; lodgify_rate_settings has minimal description); parameter descriptions are generic or incomplete (many 'Query parameters' without detail); output schemas are not documented, LLMs cannot predict return structures; tool annotations (readOnlyHint) are missing despite all tools being read-only; no pagination guidance documented despite list tools returning large result sets; input validation is present but error messages are not actionable enough to guide LLM recovery. Naming is consistent (verb_noun pattern: get_, list_) and no naming conflicts detected. The server implements modern patterns (Zod schemas, error-wrapper, input sanitizer) but lacks LLM-facing polish (output documentation, actionable error guidance, tool annotations). Average tool score: 62/100.
View daily pricing rates for properties across date ranges. This shows the actual nightly rates that would be charged for specific dates.
Retrieve comprehensive details for a specific property including configuration, amenities, room types, location information, and booking settings.
Get availability for a specific property over a period. Useful for granular checks before booking or blocking dates.
Calculate pricing for a potential booking. This tool checks availability and calculates total pricing for specified dates, room types, and guest counts.
Retrieve a messaging conversation thread including all messages, participants, and thread metadata. This is currently the only functional messaging endpoint in Lodgify API v2.
lodgify_get_quote tool has no visible description or input schema in provided source. Cannot assess LLM-facing documentation or parameter guidance. This tool is likely functional but completely undocumented for agent use.
Output schemas are not documented anywhere. LLMs cannot predict what fields to extract from responses, forcing them to infer structure from examples or failing gracefully. list_* tools return paginated results but pagination metadata (total_count, next_cursor) are not declared, preventing LLMs from planning multi-page iterations.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) present despite all 9 tools being marked Risk: READ_ONLY. Tool annotations are a current MCP spec 2026-07-28 pattern, agents should declare immutability to optimize planning and reduce safety concerns.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 51 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 48 | - | v1 |
Retrieve bookings with comprehensive filtering options. Filter by dates, status, property, guest information, and more. Returns booking details including guest info, dates, pricing, and payment status.
List all properties with optional filtering and pagination. Returns property details including names, IDs, locations, and basic configuration.
List all properties that are vacant for a date range, optionally including room details. Use this to find available inventory in one call instead of checking each property separately.
Retrieve rate configuration settings including pricing rules, modifiers, seasonal adjustments, and rate calculation parameters.
Parameter descriptions are often generic ('Query parameters', 'Required if...') without actionable detail. Examples: lodgify_list_bookings 'stayFilter' enum is well-listed but 'stayFilterDate' says 'Required when stayFilter is ArrivalDate or DepartureDate', does the LLM know the date format (YYYY-MM-DD)? lodgify_rate_settings 'params' object has description 'Query parameters for rate settings' with no field-level guidance.
Error handling in wrapToolHandler and input sanitizer is present but error messages lack recovery guidance. When an invalid date is rejected, the error message is 'Invalid from date: ...', it does NOT suggest 'Try format YYYY-MM-DD' or 'Call lodgify_get_property_availability with dates between 2024-01-01 and 2025-12-31'. Agents have no path forward.
lodgify_list_bookings accepts 'page' (default 1) and 'size' (default 50, max 50) but description does not indicate total result count or pagination behavior. If there are 10,000 bookings, how does the LLM know it needs 200+ pages? Missing 'includeCount' guidance for discovery.
lodgify_list_vacant_inventory includes debug and timeoutSeconds parameters but these are not typical agent concerns, they expose internal implementation details (timeout tuning, debug flags). Agents should not reason about timeouts or debug modes; these belong in server configuration, not tool parameters.
Tool names use Lodgify domain prefix (lodgify_*) consistently, which is good for clarity in a crowded tool namespace, but param names do NOT consistently reflect type suffixes. Examples: 'propertyId' (should 'houseId' == 'propertyId'?), 'roomTypeId', 'threadGuid'. Inconsistency (Id vs Guid vs no suffix) invites parameter type confusion when chaining tools.