What are the key innovations of Direwolf?

Direwolf features a RAM-based database, a simple low-level REST/JSON-RCP 2.0 hybrid API, an MCP-compatible data format, and a unique URI-based schema for accessing BIM entities.

Which Autodesk Revit versions are supported?

Direwolf is currently built for Autodesk Revit 2025 and up.

Direwolf is an experimental framework for BIM applications, specifically Autodesk Revit, focused on robust data extraction, manipulation, and transformation to facilitate communication with MCP-compatible AI agents.

How does Direwolf handle data and communication?

It uses a RAM-based database (Wolfden), a dedicated server (Hunter), and an MCP-compatible JSON-RCP 2.0 data format (Wolfpack) for efficient data transport and API interaction.

Is Direwolf ready for production use?

No, Direwolf is a proof-of-concept (v0.2-alpha) and is not a finalized product. It's under active development and not recommended for production environments.

Direwolf

Name: Direwolf
Author: Framebuffers

•

Enables robust data extraction, manipulation, and transformation for BIM applications, specifically Autodesk Revit, to facilitate communication with MCP-compatible AI agents.

Direwolf is an experimental, proof-of-concept data analysis framework designed for Autodesk Revit and other BIM applications. It provides a complete framework and protocol for extracting, manipulating, and transforming BIM data, enabling seamless data movement within applications, to external tools, and specifically to MCP-compatible LLM (AI) agents. Comprising a host (Direwolf), a RAM-based database (Wolfden), a server (Hunter), and a specialized data format (Wolfpack), it aims to establish a common language for BIM data across diverse sources and storage solutions, addressing contemporary and future technological challenges in the BIM ecosystem.

主要功能

01Provides a simple, low-level API with a REST/JSON-RCP 2.0 hybrid approach.

02Features a completely RAM-based database (Wolfden) and a dedicated server (Hunter).

0328 GitHub stars

04Utilizes Wolfpack, an MCP-compatible JSON-RCP 2.0 based data format.

05Includes an URI-based schema for accessing any BIM-related entity, property, or taxonomy.

06Experimental data extraction engine and MCP-based server for BIM applications.

使用案例

01Establishing a common data language for diverse BIM data sources and storage solutions.

02Facilitating data exchange between BIM software and MCP-compatible AI agents.

03Extracting, manipulating, and transforming data from Autodesk Revit models.