Tools Reference
All available MCP tools exposed by the Mendix Expert server.
Search Tools
query_mendix_knowledge
Search the curated knowledge base for Mendix topics.
@mendix-expert query_mendix_knowledge topic="microflow error handling" detail_level="detailed"
| Parameter | Type | Default | Description |
|---|---|---|---|
topic | string | required | The topic to search for |
detail_level | enum | basic | basic, detailed, or expert |
Returns: Ranked list of matching knowledge entries with relevance scores.
hybrid_search
Combined keyword + semantic search for best results.
@mendix-expert hybrid_search query="iterate over list" limit=5
| Parameter | Type | Default | Description |
|---|---|---|---|
query | string | required | Search query |
limit | number | 10 | Max results to return |
keywordOnly | boolean | false | Skip vector search |
vectorOnly | boolean | false | Skip keyword search |
Returns: Fused results with match type indicators (🎯 both, 📝 keyword, 🔮 vector).
vector_search
Pure semantic search using embeddings.
@mendix-expert vector_search query="business logic automation"
| Parameter | Type | Default | Description |
|---|---|---|---|
query | string | required | Semantic search query |
topK | number | 10 | Number of results |
minScore | number | 0.3 | Minimum similarity score |
Returns: Semantically similar entries from Pinecone.
Project Analysis
analyze_project
Analyze a Mendix .mpr file or extracted project.
@mendix-expert analyze_project project_path="D:/Projects/MyApp.mpr"
| Parameter | Type | Default | Description |
|---|---|---|---|
project_path | string | required | Path to .mpr file |
module_name | string | optional | Specific module to analyze |
entity_name | string | optional | Specific entity (requires module_name) |
Returns: Project structure with modules, entities, microflows, and pages.
analyze_theme
🎨 NEW! Deep analysis of Mendix custom themes with letter grading.
@mendix-expert analyze_theme project_path="D:/Projects/MyApp" mendix_version="11"
| Parameter | Type | Default | Description |
|---|---|---|---|
project_path | string | required | Path to Mendix project directory |
mendix_version | string | auto-detect | 10 or 11 - version of Studio Pro |
Returns:
- grade - Letter grade (A+, A, A-, B+, B, B-, C+, C, C-, D+, D, D-, F)
- score - Numeric score 0-100
- categoryScores - Individual scores for:
fileStructure- Proper theme folder organizationnaming- BEM conventions, widget prefixesvariables- Use of design tokens and SCSS variablestheming- Atlas UI integration, design system compliancemaintainability- Code organization, documentationperformance- Efficient selectors, minimal specificity
- issues - Array of problems found with severity, file, line, and suggestions
- recommendations - Prioritized list of improvements
Example Output:
{
"grade": "B+",
"score": 86,
"categoryScores": {
"fileStructure": 90,
"naming": 85,
"variables": 88,
"theming": 82,
"maintainability": 85,
"performance": 86
},
"issues": [
{
"severity": "warning",
"category": "naming",
"message": "Button class should use 'btn-' prefix",
"file": "theme/web/custom/_buttons.scss",
"line": 42,
"suggestion": "Rename .primary-button to .btn-primary"
}
],
"recommendations": [
"Consider using Atlas design tokens for colors",
"Add a _variables-custom.scss file for overrides"
]
}
Knowledge Management
add_to_knowledge_base
Add new knowledge to make the system smarter.
@mendix-expert add_to_knowledge_base knowledge_file="best-practices" category="microflows" content="{...}" source="docs.mendix.com"
| Parameter | Type | Default | Description |
|---|---|---|---|
knowledge_file | enum | required | Target file (see below) |
category | string | optional | Category within file |
content | string | required | JSON string of knowledge entry |
source | string | required | Where this info came from |
verified | boolean | false | Whether manually verified |
Knowledge Files:
best-practices- Patterns and guidelinesstudio-pro- Studio Pro featuresmodel-sdk- Model SDK programmingplatform-sdk- Platform SDK operationstroubleshooting- Issues and solutionsadvanced-patterns- Complex techniquesperformance-guide- Optimizationsecurity-guide- Security practices
Returns: Entry ID, quality score, confirmation of re-indexing.
get_best_practice
Get recommendations for specific scenarios.
@mendix-expert get_best_practice scenario="error handling in microflows"
| Parameter | Type | Default | Description |
|---|---|---|---|
scenario | string | required | The scenario to get advice for |
Harvesting
harvest
Crawl Mendix documentation for fresh knowledge.
@mendix-expert harvest sources=["releaseNotes", "mxsdk"] dryRun=false
| Parameter | Type | Default | Description |
|---|---|---|---|
sources | array | all | Specific sources to harvest |
dryRun | boolean | false | Preview without saving |
verbose | boolean | true | Show detailed progress |
Available Sources:
releaseNotes- Studio Pro 10.x, 11.x release notesrefGuide- Reference guide sectionshowTo- How-to guidesstudioProGuide- Studio Pro guideapidocs- API documentationmxsdk- SDK documentation
harvest_status
Check harvest schedule and history.
@mendix-expert harvest_status
Returns: Last harvest date, next scheduled, total harvests, available sources.
Vector Search Management
vector_status
Check Pinecone index and embedding status.
@mendix-expert vector_status
Returns:
- Index status (ready/not initialized)
- Vector count
- Embedding mode (azure-openai/openai/local)
- Query cache stats (hit rate)
reindex_vectors
Re-index all knowledge for vector search.
@mendix-expert reindex_vectors
Use this after:
- Major knowledge base updates
- Changing embedding provider
- Suspected index corruption
Sync & Maintenance
sync_mcp_server
Sync knowledge with GitHub repository.
@mendix-expert sync_mcp_server action="pull"
| Parameter | Type | Default | Description |
|---|---|---|---|
action | enum | pull | pull, push, or status |
Research Mode
beast_mode
Get the exhaustive research protocol prompt.
@mendix-expert beast_mode format="full"
| Parameter | Type | Default | Description |
|---|---|---|---|
format | enum | full | full, brief, or instructions |
Returns: Copy-paste ready prompt that enables aggressive research mode in any AI chat.
hello
Welcome screen with status overview.
@mendix-expert hello
Returns: Server status, knowledge stats, vector status, quick examples.
Analytics Tools
get_usage_analytics
View usage statistics, popular topics, and trends.
@mendix-expert get_usage_analytics period="week" include_trends=true
| Parameter | Type | Default | Description |
|---|---|---|---|
period | enum | all | day, week, month, or all |
include_trends | boolean | true | Include hourly/daily trends |
Returns: Tool usage counts, popular topics, query patterns, and trend data.
Tool Usage Patterns
Finding Information
# Start broad, then narrow
@mendix-expert query_mendix_knowledge topic="security"
@mendix-expert hybrid_search query="user role xpath constraints"
Learning Mode
# Research → Discover → Save
@mendix-expert beast_mode format="full"
# [Use prompt to research topic]
@mendix-expert add_to_knowledge_base knowledge_file="best-practices" ...
Project Work
# Understand project structure
@mendix-expert analyze_project project_path="D:/MyApp.mpr"
# Find relevant patterns
@mendix-expert hybrid_search query="domain model for [your entities]"
Check Your Usage
# See what you've been searching for
@mendix-expert get_usage_analytics period="week"