RACCORDO is an Autodesk® Revit® add-in that uses AI to automate the classification of Revit element types with OmniClass Table 23 Keynote values and CSI MasterFormat 2020 Assembly Codes — eliminating the manual lookup and entry work that BIM managers face when maintaining consistent parameter data across large shared parameter libraries.

On most projects, Revit models contain dozens to hundreds of element types — wall types, floor types, structural framing, MEP equipment, doors, windows, and more — that should carry standardised classification codes for BIM compliance, asset management, and downstream data exchange. Without a tool like Raccordo, a BIM manager must look up each type in the OmniClass or MasterFormat tables, copy the code, and paste it into the correct parameter in Revit — a repetitive process that is slow, error-prone, and frequently skipped.

Raccordo replaces this process with a three-step workflow. Load Knowledge Base opens a viewer of the active classification knowledge base — a curated JSON file mapping element name patterns to OmniClass and MasterFormat values. Suggest Parameter Values opens an element type browser where the user selects the types to classify. Raccordo then runs a local RAG (Retrieval-Augmented Generation) pipeline to find the closest knowledge base matches, and sends those matches together with the element type name to the configured AI provider for a final suggested code. The user reviews and confirms the suggestions, and Raccordo writes the accepted values to the Revit model in a single batch transaction. Audit Parameter Coverage generates a compliance report showing which element types have complete, partial, or missing classification data — giving the BIM manager an instant view of the model's classification state.

Key Features:

• AI-powered classification: choose your AI provider — Anthropic Claude, OpenAI, or a local Ollama model — to suggest OmniClass T23 Keynote and MasterFormat Assembly Code values for each selected element type

• Local-first option: run entirely offline using Ollama — no data leaves your machine, no API key required

• RAG pipeline: local keyword retrieval finds the closest knowledge base matches before calling the AI, improving suggestion accuracy and reducing API cost

• Batch apply: confirm suggestions across multiple element types and write all values in a single Revit transaction

• Audit report: Complete / Partial / Missing compliance view across all 21 classifiable element categories

• Knowledge base driven: swap or expand the KB JSON file to match project-specific or organisational classification standards

• Revit 2023 -2027 supported.

• Paid — one-time purchase.

Note: The Suggest command requires an AI provider to be configured (Anthropic Claude, OpenAI, or Ollama). The Audit command always works offline. Suggestions should be reviewed before applying — AI outputs are not guaranteed to be correct.

Note: This app uses a custom installer (and not the standard App Store installer).