How do I use HOOPS AI MCP Server?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@HOOPS AI MCP Server Display the 3D CAD file 'C:\temp\part.step'." That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

HOOPS AI MCP Server

by toshi-bata

Overview Schema Related Servers Score Discussions

Python

Local

HOOPS AI MCP Server

An MCP (Model Context Protocol) server that bridges Claude Desktop to the HOOPS AI WebAPI.
With this server registered in Claude Desktop, users can perform 3D CAD analysis through natural language — no code required.
See the root README for an overview of the full platform.

Prerequisites

uv installed on the Claude Desktop machine (Claude Desktop uses uv to launch the MCP server process)
The WebAPI server running and accessible (default: http://127.0.0.1:8000)
→ See webapi/README.md for setup instructions

Related MCP server: Cinema4D MCP Server

Setup

Register the MCP server in Claude Desktop

Open Claude Desktop
Go to Settings → Developer → Edit Config
This opens claude_desktop_config.json. Add the following entry under mcpServers:

{
  "mcpServers": {
    "hoops-ai": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "C:\\path\\to\\HOOPS_AI-MCP\\mcp_server",
        "server.py"
      ]
    }
  }
}

Replace C:\\path\\to\\HOOPS_AI-MCP with the actual path where you cloned this repository.

Troubleshooting — uv not found: Claude Desktop launches with a limited PATH and may fail to find uv even if it works in your terminal.
If the MCP server does not appear in Claude Desktop, use the full path to uv.exe instead of "uv":
where.exe uv   # find the full path, e.g. C:\Users\<you>\.local\bin\uv.exe
Then update "command" in the config:
"command": "C:\\Users\\<you>\\.local\\bin\\uv.exe"

Same machine (default):

No additional configuration is needed.
The MCP server defaults to http://127.0.0.1:8000, so if the WebAPI server is running on the same machine, the basic config above works as-is.

When the WebAPI server is on a different machine (client-server setup):

Add "env": {"HOOPS_WEBAPI_URL": "..."} to the config — no system environment variable is needed.
Claude Desktop passes this value to the MCP server process automatically:

{
  "mcpServers": {
    "hoops-ai": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "C:\\path\\to\\HOOPS_AI-MCP\\mcp_server",
        "server.py"
      ],
      "env": {
        "HOOPS_WEBAPI_URL": "http://192.168.0.6:8000"
      }
    }
  }
}

Replace 192.168.0.6 with the actual IP address of the machine running the WebAPI server.
This is the only configuration change needed on the client machine.

Save the file and restart Claude Desktop.

(Optional) Import the Skill file

The skills/ folder contains a .skill file that defines the expected behaviors and best-practice instructions for using this MCP server with Claude Desktop.
Importing it gives Claude a consistent baseline for how to invoke the HOOPS AI tools — you can then customize the skill to match your own workflow.

How to import:

In Claude Desktop, go to Settings → Skills
Click Import Skill and select skills/hoops-ai-tool-tips.skill from this repository
Review and edit the skill content as needed for your use case

Available MCP Tools

Claude Desktop can call these 33 tools using natural language:

File Management

Tool	Description
`upload_cad_model`	Upload a local CAD file to the server. Returns `file_id`, `filename`, and `already_existed`. Pass `file_id` to other tools to avoid re-uploading.
`open_cad_viewer`	Open a CAD file in the interactive 3D browser viewer. Returns `viewer_url` and `image_url` (PNG preview).
`terminate_CAD_viewer`	Terminate the last active viewer, or all viewers (`terminate_all=True`).

B-Rep Analysis

Tool	Description
`get_brep_adjacency_graph`	Build a face adjacency graph from a CAD file. Returns graph data (nodes, edges, counts) and `image_url` (PNG visualization URL).
`get_brep_attributes`	Extract raw per-face and per-edge attributes (types, areas, lengths, dihedral angles, etc.) from a CAD file. For individual-entry questions, not counting.
`get_brep_type_counts`	Return face and edge counts grouped by type, aggregated server-side. Use for any counting question ("how many faces", "faces by type", totals).

Manufacturing Feature Recognition (MFR)

Tool	Description
`get_MFR_table_of_contents`	Get a summary of the MFR dataset.
`get_MFR_labels_description`	List all MFR label IDs, feature names, and descriptions.
`search_MFR_files`	Find CAD files in the MFR dataset that contain a given manufacturing feature.
`get_MFR_file_thumbnail`	Return the URL of the thumbnail PNG for a given dataset file ID.
`run_MFR_inference`	Run MFR inference on a CAD file. Returns predictions, probabilities, `viewer_url`, and `image_url`.

Shape Similarity Search

Tool	Description
`search_similar_shapes`	Find the top-k most similar parts using HOOPS Embeddings and a FAISS index. Returns match IDs, similarity scores, and `image_url`.
`get_similar_part_image`	Return the URL of the pre-generated PNG thumbnail for a part filename returned by `search_similar_shapes`.
`get_similar_search_index_info`	Return metadata about the loaded FAISS index: status, entry count, embedding model name, vector dimension, file path, last-modified timestamp, and auxiliary metadata. Read-only.
`embed_cad_shape`	Compute the shape embedding for a single CAD part (no FAISS index or training required). Returns `file_id`, `filename`, `dim`, `model_name`, `num_bodies`, and `cached`. Embeddings are cached server-side for fast repeated calls.
`get_embedding_settings`	Return the server-wide active embedding model (`'signal'` or `'default'`). Used by `compare_cad_shapes`, `generate_shape_space_map`, and `create_similarity_index`.
`set_embedding_model`	Set the server-wide active embedding model: `'signal'` (HOOPS AI SIGNAL model, default) or `'default'` (1M model). Affects all subsequent compare/map/index-create calls. Existing indexes are unaffected.
`compare_cad_shapes`	Compute pairwise cosine-similarity scores for 2+ CAD parts (no FAISS index or training required). Returns an N×N similarity matrix, a ranked pair list, and per-file error details. Accepts local paths, existing `file_id`s, and/or a ZIP path. ZIP files are processed server-side (no large upload). Uses the server-wide active model (default: `'signal'`).

Named Similarity Index Management

Tool	Description
`create_similarity_index`	Create a new empty named index. The embeddings model is taken from the server-wide setting (`set_embedding_model`; default `'signal'`). Persists across server restarts. Returns `name`, `count` (0), `dim`, and `model`. Raises 409 if the name already exists.
`list_similarity_indexes`	List all similarity indexes including the built-in read-only `default` index. Each entry contains `name`, `count`, `last_modified`, `is_readonly`, and `model`.
`add_to_similarity_index`	Register CAD parts in a named index. Accepts local paths, `file_id`s, and/or a ZIP path. ZIP files are processed server-side (no large upload). The embedder is always the one recorded in the index at creation time (`model.json`). Returns `added`, `updated`, `index_count`, and `errors`.
`search_similarity_index`	Search a named index for the top-k most similar parts to a query CAD file. The correct embedder is selected automatically based on the index model. Returns hits with `id`, `score`, `metadata`, and an `image_url` result-grid PNG.
`remove_from_similarity_index`	Remove specific parts (by `file_id`) from a named index. Returns `removed` count and `index_count` remaining.
`delete_similarity_index`	Permanently delete a named index and all its stored data. Irreversible. Raises 403 for the built-in `default` index.

Shape Space Map

Tool	Description
`generate_shape_space_map`	Generate a Shape Embeddings Map and poll internally until complete (up to 30 min). Accepts local paths, `file_id`s, and/or a ZIP path. ZIP files are processed server-side. Returns `map_id`, `viewer_url`, per-part MDS `position`, similarity `matrix`, and MDS `stress` directly — no polling needed in normal use.
`get_shape_space_map_result`	Fallback poll for a map job started by `generate_shape_space_map`. Only needed if `generate_shape_space_map` returned `status: "processing"` because the 30-minute deadline was exceeded. Returns `status` (`"processing"` / `"done"` / `"failed"`); when `"done"`, the full result is included.
`query_shape_space_map`	Project a query CAD part onto an existing Shape Space Map (highlighted in magenta). Set `persist=true` to permanently add the query part to the original map. Returns `overlay_map_id`, `viewer_url`, and `nearest_parts`.

Part Classification

Tool	Description
`run_part_classification_inference`	Run Part Classification inference on a CAD file. Returns the top-k predicted part classes with confidence scores (1–45 classes).
`get_part_classification_labels`	Return the full 45-class part label dictionary with IDs and descriptions.
`get_part_classification_table_of_contents`	Get a summary of the Part Classification dataset including available groups.
`get_part_classification_label_distribution`	Return per-class file count distribution across the Part Classification training dataset.
`get_part_classification_files`	Return the list of file IDs in the dataset that belong to a given part class (`label_id` 0–44).
`get_part_classification_preview`	Return a URL to a PNG thumbnail grid for a given part class (`label_id`, `k`, `grid_cols`).

Context Layer (Missing Metadata Prediction)

Tool	Description
`predict_context`	Predict missing metadata (material, cost, lead time, etc.) for a query CAD part using similar parts' history. Step 3 of a 3-step workflow: (a) `search_similar_shapes` → hits, (b) PLM/ERP tool → contexts, (c) this tool → predictions. Returns `predictions` with `value`, `confidence`, and `status` per key.

Example Usage in Claude Desktop

Once the MCP server is registered and the WebAPI server is running, you can chat with Claude:

What HOOPS AI tools are available?

"C:\temp\helloworld.stp" — please display this 3D CAD file.

この部品の材料とコストを、似ている部品の実績から予測して

Predict the missing material and cost for this part based on similar parts' history.

"C:\temp\Flange287.stp" — show this model and give me its B-Rep information.

Tell me about the manufacturing feature recognition dataset.

"C:\temp\nist_ftc_06_asme1_rd_sw1802.SLDPRT" — run manufacturing feature recognition and colorize by feature type.

"C:\temp\idler_sprocket.step" — search for similar parts to this component.

この2つのSTEPファイルはどれくらい似ている？ C:\temp\partA.stp と C:\temp\partB.stp

Compare these three parts and tell me which two are most similar:
C:\temp\bracket_v1.step, C:\temp\bracket_v2.step, C:\temp\bracket_v3.step

ZIPに入っているCADファイルの類似度マトリクスを出して。ファイルは C:\temp\parts.zip

Compute the shape embedding for C:\temp\flange.stp and tell me the embedding dimension and model name.

Create a named index called "my-brackets" and register all STEP files in C:\temp\brackets.zip.

Search my-brackets index for parts similar to C:\temp\new_bracket.step and show me the top 5 results.

Generate a 3D shape space map for these parts: C:\temp\partA.stp, C:\temp\partB.stp, C:\temp\partC.stp

ZIPに入っている全CADファイルの形状空間マップを生成して。ファイルは C:\temp\parts.zip

Query the shape map <map_id> with C:\temp\new_part.step and show me the nearest parts.

Note — ZIP file processing: When a ZIP path is passed to compare_cad_shapes, add_to_similarity_index, or generate_shape_space_map, the WebAPI server reads the file directly from the given path. This requires the MCP client and the WebAPI server to be on the same machine (the default local setup). For remote setups (WebAPI on a separate host), use upload_cad_model to upload individual files first, then pass their file_ids.

This server cannot be installed

license - not found

quality - not tested

maintenance

How are these scores calculated?

Maintenance

–Maintainers

–Response time

–Release cycle

–Releases (12mo)

Commit activity

Resources

GitHub Repository

Need Help?

Related Servers

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)
By Om-Shree-0709 on July 25, 2026.
mcp
Agent Identity
OAuth 2.1
Your AI Chatbot Just Exposed Your CEO's Salary to an Intern
By Om-Shree-0709 on July 2, 2026.
Agent Identity
MCP Security
OAuth Delegation
Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)
By Om-Shree-0709 on June 30, 2026.
Agentic Ai
Prompt Injection
WebAssembly

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/toshi-bata/HOOPS_AI-MCPServer'

If you have feedback or need assistance with the MCP directory API, please join our Discord server