SketchUp Agent Harness is an open-source framework designed to connect agent command-line interfaces, an MCP server, a SketchUp Ruby bridge, and structured design models into a verifiable design workflow. The project addresses the limitations of typical AI design tools, which often stop at generating images or executing a few software steps, by enabling a workflow where design intent, source materials, and project state are explicitly recorded, checked, and repairable. json file serving as the central record of project state, including spaces, dimensions, components, rules, assumptions, and execution plans.
The harness allows designers to express their goals in natural language, which are then processed by agent CLIs such as Claude and Codex. These agents interact with runtime skills, the MCP server, and the SketchUp Ruby bridge to translate intent into executable and auditable project actions within SketchUp. Key features include initializing isolated project directories, installing the SketchUp Ruby bridge, integrating Claude and Codex runtime skills, generating basic spaces and bathroom layouts from natural language, and importing a range of input formats such as DWG, DXF, PDF, images, scans, and photos into editable models. The tool preserves structured design data, source evidence, screenshots, version history, project rules, imports, and assets, and creates inspectable execution traces before SketchUp operations are performed. It writes entity IDs, execution status, and metadata back into the project’s structured truth for ongoing validation and repair.
SketchUp Agent Harness is intended for designers and developers exploring how natural-language-driven agents can reliably control SketchUp and organize the design process into a structured, checkable workflow. The project is not positioned as a mature commercial design platform but rather as an early-stage harness for experimentation and workflow development. It is delivered as an open-source package, with installation steps involving the harness itself, the SketchUp Ruby bridge, and the use of agent CLIs within a clean project directory. The project’s documentation and repository are publicly available, and its open-source status is explicitly stated.
By focusing on verifiable protocols, structured project state, and auditable execution traces, SketchUp Agent Harness aims to support workflows where design changes are explicit, traceable, and repairable, allowing designers to concentrate on higher-level intent and decision-making.
SketchUp Agent Harness sits in PulseGate's Frameworks & SDKs category. It focuses on enabling design agents to interact with SketchUp through a verifiable, structured workflow and protocol. SketchUp Agent Harness is an open-source project aimed at AI design agent developers. The project is open source (MIT). SketchUp Agent Harness is available on the web, the command line, and API, and it can be self-hosted.
Marlin Bian builds and maintains SketchUp Agent Harness, and the product first shipped in 2026. The project is developed in the open on GitHub with 138 commits in the last 90 days. Among its 5 catalogued features are MCP server integration, SketchUp Ruby bridge, and design model verification. The interface is available in English and Chinese. It exposes integrations via an MCP server. SketchUp Agent Harness is currently in beta.
Latest indexed changes and source events
Other apps tracked under the same category.