Volume: 125.5 cm³
', 'width': 300, 'height': 150, 'tool_unlock_token': '1d9bf6a0' } }) # Open documentation in browser mcp.call('browser', { 'input': { 'operation': 'navigate', 'url': 'https://help.autodesk.com/view/fusion360/', 'tool_unlock_token': 'e5076d' } }) ``` -------------------------------- ### Atomic Commits for Patch Generation Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/CONTRIBUTING.md This example demonstrates a workflow for making small, logical changes and committing them individually. Atomic commits are crucial for generating clean, reviewable patch files. Each commit should have a clear message explaining the change. ```bash # Example of a good commit workflow # (edit file1.js, file2.js) git add . git commit -m "feat: Add geolocation tool foundation" # (edit file3.js) git add . git commit -m "fix: Correct error handling in geolocation response" ``` -------------------------------- ### SSE Connection Establishment Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/INTEGRATION_GUIDE.md This describes the Server-Sent Events (SSE) connection flow. It involves making a GET request to the SSE endpoint, receiving a session ID, storing the message endpoint, and starting a reader thread. ```text GET /sse → Receive session_id → Store message endpoint → Start reader thread ``` -------------------------------- ### MCP Client Startup Log Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/IMPLEMENTATION_SUMMARY.md This log details the successful startup sequence of the MCP client when MCP_AUTO_CONNECT is set to True. It shows the steps from finding the manifest to registering with the MCP server and establishing an SSE connection. ```text ============================================================ MCP Client Connection Starting ============================================================ Step 1: Finding native messaging manifest... [OK] Found manifest: C:\Users\...\com.aurafriday.shim.json Step 2: Reading manifest... [OK] Manifest loaded Step 3: Discovering MCP server endpoint... [OK] Server configuration received Step 4: Extracting server URL and auth... [OK] Server URL: https://127-0-0-1.local.aurafriday.com:XXXXX/sse Step 5: Connecting to SSE endpoint... [OK] SSE Connected! Session ID: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx Step 6: Checking for remote tool... [OK] Remote tool found Step 7: Registering fusion360 with MCP server... ============================================================ [SUCCESS] fusion360 registered successfully! Listening for reverse tool calls in background... ============================================================ ``` -------------------------------- ### MCPClient Connection Sequence Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/ARCHITECTURE.md Details the step-by-step connection process initiated by the MCPClient to establish communication with the MCP Server. This flow involves finding manifests, running binaries, and registering tools. ```bash # User clicks 'Connect' # Fusion 360 creates a new MCPClient # MCPClient finds manifest # MCPClient reads manifest # MCPClient runs binary (with JSON config) # MCPClient connects to SSE # MCPClient registers tool # MCPClient starts listener # Fusion 360 receives connection confirmation ``` -------------------------------- ### Minimal Fusion 360 API Example: Extruded Text (Python) Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/extrude_text.md A complete, copy-pasteable Python script for Fusion 360 that demonstrates the entire process: getting the design, creating a sketch, adding text, and extruding it. This example is designed to be robust against recent API changes. ```Python # Always run inside Fusion's Python environment import adsk.core import adsk.fusion def run(context): ui = None try: app = adsk.core.Application.get() ui = app.userInterface design = adsk.fusion.Design.cast(app.activeProduct) if not design: ui.messageBox('No active Fusion design found.', 'Error') return root_comp = design.rootComponent # 1. Create a new sketch on the XY plane xy_plane = root_comp.xYConstructionPlane sketches = root_comp.sketches sketch = sketches.add(xy_plane) # 2. Create sketch text input text_input = None if hasattr(sketch.sketchTexts, 'createInput3'): text_input = sketch.sketchTexts.createInput3() else: # Fallback for older versions, height in cm text_input = sketch.sketchTexts.createInput2(adsk.core.ValueInput.createByReal(2.0)) # 2 cm height text_input.fontName = "Arial" text_input.text = "API Text" # Example: set as multi-line with center alignment text_input.setAsMultiLine() text_input.horizontalAlignment = adsk.fusion.HorizontalAlignments.CenterHorizontalAlign text_input.verticalAlignment = adsk.fusion.VerticalAlignments.CenterVerticalAlign # 3. Add the sketch text entity sketch_text = sketch.sketchTexts.add(text_input) # Position the text (example: move it after creation) # Note: Direct manipulation of SketchText position is limited. # Often, you'd define the position during input creation or transform the sketch. # For simplicity, we'll assume default placement is acceptable or transform later. # 4. Extrude the sketch text features = root_comp.features extrude_features = features.extrudeFeatures # Use the simplest extrude method distance_input = adsk.core.ValueInput.createByReal(1.0) # 1 cm extrusion extrude_input = extrude_features.addSimple(sketch_text, distance_input, adsk.fusion.FeatureOperations.NewBodyFeatureOperation) ui.messageBox('Successfully created extruded text.') except Exception as e: if ui: ui.messageBox(f'Failed: {str(e)}', 'Error') ``` -------------------------------- ### Object Construction Examples in Python Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/README.md Illustrates how to construct various Fusion 360 objects using dictionaries, including points, vectors, value inputs with different formats, and empty collections. It also shows how to reference stored objects and use enums as integers. ```python { "type": "Point3D", "x": 5, "y": 10, "z": 0 } { "type": "Vector3D", "x": 1, "y": 0, "z": 0 } { "type": "ValueInput", "value": 2.5 } { "type": "ValueInput", "expression": "2.5 cm" } { "type": "ObjectCollection" } "$my_sketch.originPoint" 0 ``` -------------------------------- ### AI Agent Request to Control Fusion 360 Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/INTEGRATION_GUIDE.md This shows an example of an AI agent's request to control Fusion 360, specifically asking to create a 10cm cube. ```text "Use the fusion360 tool to create a 10cm cube" ``` -------------------------------- ### Set Extrusion Start Extent with Offset in Python Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/best_practices.md Precisely positions the start of an extrusion operation using a base plane and a real-valued offset. This method is recommended over move operations for reliable positioning. ```python extInput.startExtent = adsk.fusion.FromEntityStartDefinition.create( basePlane, adsk.core.ValueInput.createByReal(height) ) ``` -------------------------------- ### Online Documentation Fetch Source: https://context7.com/aurafriday/fusion-360-mcp-server/llms.txt Fetches rich documentation from Autodesk's cloudhelp pages, including parameter tables, return types, and working code samples. ```APIDOC ## POST /fusion360/execute (Online Doc Fetch) ### Description Fetches detailed documentation from Autodesk's cloudhelp pages. ### Method POST ### Endpoint /fusion360/execute ### Parameters #### Request Body - **operation** (string) - Required - Set to "get_online_documentation". - **class_name** (string) - Required - The name of the class to get documentation for. - **member_name** (string) - Optional - The name of the member (method or property) to get documentation for. ### Request Example (Class Level) ```json { "fusion360.execute": { "operation": "get_online_documentation", "class_name": "ExtrudeFeatures" } } ``` ### Request Example (Member Level) ```json { "fusion360.execute": { "operation": "get_online_documentation", "class_name": "ExtrudeFeatures", "member_name": "createInput" } } ``` ### Response #### Success Response (200) - **documentation** (object) - An object containing detailed documentation. - **url** (string) - The URL of the online documentation. - **class_name** (string) - The name of the class. - **member_name** (string) - The name of the member (if specified). - **description** (string) - A description of the class or member. - **parameters** (array) - An array of parameter objects (if applicable). - **return_type** (string) - The return type of the member (if applicable). - **return_description** (string) - A description of the return value (if applicable). - **samples** (array) - An array of code sample objects (if applicable). #### Response Example ```json { "url": "https://help.autodesk.com/cloudhelp/ENU/Fusion-360-API/files/ExtrudeFeatures_createInput.htm", "class_name": "ExtrudeFeatures", "member_name": "createInput", "description": "Creates an ExtrudeFeatureInput object...", "parameters": [ {"name": "profile", "type": "Base", "description": "The profile to extrude..."}, {"name": "operation", "type": "FeatureOperations", "description": "The feature operation..."} ], "return_type": "ExtrudeFeatureInput", "return_description": "The newly created ExtrudeFeatureInput object", "samples": [ {"name": "SimpleExtrude", "description": "Creates a simple extrusion", "url": "..."} ] } ``` ``` -------------------------------- ### Fusion 360 Python Tool Handler Example Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/INTEGRATION_SUMMARY.md An example implementation of a tool handler function in Python for Fusion 360. This function processes incoming commands from the MCP server, currently echoing back a confirmation, with a placeholder for future API calls. ```python import adsk.core import adsk.fusion def fusion_tool_handler(call_data): command = call_data['params']['arguments']['command'] params = call_data['params']['arguments'].get('parameters', {}) if command == 'create_box': app = adsk.core.Application.get() design = app.activeProduct rootComp = design.rootComponent # Create sketch sketches = rootComp.sketches xyPlane = rootComp.xYConstructionPlane sketch = sketches.add(xyPlane) # Draw rectangle length = params.get('length', 10) width = params.get('width', 10) lines = sketch.sketchCurves.sketchLines rect = lines.addTwoPointRectangle( adsk.core.Point3D.create(0, 0, 0), adsk.core.Point3D.create(length, width, 0) ) # Extrude height = params.get('height', 10) prof = sketch.profiles[0] extrudes = rootComp.features.extrudeFeatures extInput = extrudes.createInput(prof, adsk.fusion.FeatureOperations.NewBodyFeatureOperation) distance = adsk.core.ValueInput.createByReal(height) extInput.setDistanceExtent(False, distance) extrude = extrudes.add(extInput) return { "content": [{ "type": "text", "text": f"Created {length}x{width}x{height}cm box successfully" }], "isError": False } else: return { "content": [{ "type": "text", "text": f"Command '{command}' not yet implemented." }], "isError": True } ``` -------------------------------- ### Fetch Online Documentation for a Class Source: https://context7.com/aurafriday/fusion-360-mcp-server/llms.txt Retrieves detailed documentation for a specific class from Autodesk's cloudhelp pages. This includes parameter tables, return types, and code samples. ```python fusion360.execute({ "operation": "get_online_documentation", "class_name": "ExtrudeFeatures" }) ``` -------------------------------- ### Get Fusion Design and Root Component (Python) Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/extrude_text.md Retrieves the active Fusion 360 design and its root component. This is the initial step for most API operations that modify the design. It involves getting the application instance and casting it to a Design object. ```Python import adsk.core import adsk.fusion def get_design_and_root_component(): app = adsk.core.Application.get() design = adsk.fusion.Design.cast(app.activeProduct) if not design: raise Exception("No active Fusion design found.") root_comp = design.rootComponent return app, design, root_comp ``` -------------------------------- ### Create and Extrude 3D Text in Fusion 360 Source: https://context7.com/aurafriday/fusion-360-mcp-server/llms.txt This example demonstrates a multi-step process to create and extrude 3D text within Fusion 360. It involves creating a sketch, defining text input with dimensions, setting text properties like alignment, adding the text to the sketch, and finally extruding it. Each step utilizes the generic API call mechanism via `fusion360.execute`. ```python # Step 1: Create sketch on XY plane fusion360.execute({ "api_path": "design.rootComponent.sketches.add", "args": ["design.rootComponent.xYConstructionPlane"], "store_as": "text_sketch" }) # Step 2: Create text input (text, height in cm) fusion360.execute({ "api_path": "$text_sketch.sketchTexts.createInput2", "args": ["'Fusion Rocks!'", 1.5], "store_as": "text_input" }) # Step 3: Create corner points for text bounding box fusion360.execute({ "api_path": "adsk.core.Point3D.create", "args": [0, 0, 0], "store_as": "corner1" }) fusion360.execute({ "api_path": "adsk.core.Point3D.create", "args": [15, 3, 0], "store_as": "corner2" }) # Step 4: Configure text as multi-line (hAlign: 1=center, vAlign: 1=middle) fusion360.execute({ "api_path": "$text_input.setAsMultiLine", "args": ["$corner1", "$corner2", 1, 1, 0.0] }) # Step 5: Add text to sketch fusion360.execute({ "api_path": "$text_sketch.sketchTexts.add", "args": ["$text_input"], "store_as": "text_object" }) ``` -------------------------------- ### API Documentation Operations in Python Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/README.md Outlines the operations for retrieving API documentation through introspection, fetching online documentation, and accessing design best practices. Parameters for searching and filtering documentation are included. ```python { "operation": "get_api_documentation", "search_term": "ExtrudeFeature", "category": "class_name", "max_results": 3 } { "operation": "get_online_documentation", "class_name": "ExtrudeFeatures", "member_name": "createInput" } { "operation": "get_best_practices" } ``` -------------------------------- ### Implement Core Fusion 360 Command: Create Box (Python) Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/INTEGRATION_GUIDE.md This Python snippet demonstrates the initial steps for implementing a 'create_box' command within Fusion 360. It shows how to access the active design, create a new sketch on the XY plane, and begin drawing a rectangle using provided parameters for length and width. Further implementation is required to complete the drawing and extrusion. ```python if command == 'create_box': # Get active design design = app.activeProduct rootComp = design.rootComponent # Create new sketch sketches = rootComp.sketches xyPlane = rootComp.xYConstructionPlane sketch = sketches.add(xyPlane) # Draw rectangle length = params.get('length', 10) width = params.get('width', 10) ... ``` -------------------------------- ### Python Command Registry Update Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/INTEGRATION_GUIDE.md This snippet shows an update to the command registry in `commands/__init__.py`. It imports the `mcp_connection_handler` and adds it to the list of available commands, specifically for MCP connection. ```python from .mcpConnect import mcp_connection_handler as mcpConnect commands = [ mcpConnect, # MCP connection command (NEW) commandDialog, paletteShow, paletteSend ] ``` -------------------------------- ### Object Construction Logic Source: https://github.com/aurafriday/fusion-360-mcp-server/blob/main/docs/GENERIC_API_DESIGN.md Demonstrates the logic for constructing Fusion 360 objects dynamically based on input JSON. It shows how the system detects the object type, finds the corresponding class, checks for static creation methods, extracts parameters, and calls the appropriate method. ```text Input: {"type": "Point3D", "x": 10, "y": 5, "z": 0} ↓ 1. Detect "type" field → constructor mode ↓ 2. Find class: adsk.core.Point3D ↓ 3. Check for .create() static method → YES ↓ 4. Extract params: x=10, y=5, z=0 ↓ 5. Call: Point3D.create(10, 5, 0) ↓ 6. Return: