editor
Open-source, local-first 3D architectural editor built with WebGPU and native MCP AI agent workflows.
Build, view, and inspect 3D building models using React Three Fiber, WebGPU, local CLI, and MCP servers.
Pascal Editor is an open-source 3D building design tool tailored for local-first modeling and automated spatial agent workflows. It connects browser-based rendering directly to AI tools like Claude Code or Codex using Model Context Protocol (MCP) servers. It is built for AEC developers, architects, and researchers creating automated design systems.
What it is
The application renders detailed 3D building assemblies, floorplans, and mechanical systems using WebGPU and React Three Fiber. Its background CLI runner creates an MCP connection, allowing external AI agents to inspect scene structures, check spatial clearances, and generate model elements. Project data is saved locally to SQLite databases or hosted through a single Docker container.
Who it's for
This editor suits computational designers, AEC software engineers, and workflow developers building AI-assisted spatial applications. It does not suit traditional non-technical architects expecting an out-of-the-box drafting suite with pre-built BIM object libraries or direct DWG/IFC export tools without technical configuration.
Key features
WebGPU 3D Engine
Renders complex architectural structures and interior systems directly in modern web browsers using React Three Fiber.
Model Context Protocol Integration
Exposes model state and command execution to external AI assistants through a local CLI background service.
Local SQLite Storage
Persists project databases directly on your machine without requiring mandatory cloud accounts or external database servers.
Pre-Built Spatial Skills
Provides executable skill packages for automated 3D scene editing and furniture fit evaluation.
Up and running in 3 steps
Our honest verdict
What's great
- Built-in MCP server integration lets AI agents inspect and edit 3D building models programmatically.
- Local-first architecture keeps spatial datasets under local control inside standard SQLite files.
- Ships pre-packaged agent workflows for spatial verification and furniture footprint evaluation.
- Runs on demand via a single npx terminal command or deploys as a standard Docker image.
Where it falls short
- Requires managing Node.js environment paths and background process configurations for AI agent linking.
- Local service instance constraints allow only one active agent client per local process without custom environment isolation.
- Advanced features demonstrated in the hosted demo are not yet fully available in the open-source repository release.
Pricing & license
Pascal Editor is completely free and open source under the MIT license. You can run it locally or self-host it on your own server infrastructure at no cost. Using optional hosted cloud sync or API features requires setting up credentials on the developer platform.
Open remote AI-training roles
767 open remote AI-training roles, updated November 14, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
How does Pascal Editor connect to AI assistants like Claude Code?
It launches a background service that exposes a local Model Context Protocol (MCP) endpoint, which agent plugins connect to directly without uploading project files.
Where is my project data stored when self-hosting?
All project models and configurations are saved into an SQLite database file located at `~/.pascal/data/pascal.db` or inside the mounted `/data` Docker volume.
