BIMserver vs Speckle

BIMserver stores and manages building-project information as IFC data, using structured model objects rather than treating the project as a collection of files. Users can check, version, merge, query, and filter models, arrange project structures, and generate IFC output on demand. Multi-user support lets people work on separate parts of a dataset, with notifications when a model or section changes. BIMserver is intended to serve as an IFC backend for other software, and developers can add plugins, services, and clients or embed it in their products. The listed service interfaces are SOAP, JSON, and Protocol Buffers. It is open-source and self-managed, with a free plan under the AGPL-3.0 license; some plugins and libraries use different licenses. It supports Linux, macOS, Windows, web, API, and self-hosted use. The project wiki points to GitHub Releases for downloads and includes an Ubuntu installation guide. For help, maintainers point users to the wiki, FAQ, issue tracker, and contributors, and list an email contact. The project says its open-source community has been active since 2013.

Speckle is a design data platform for architecture, engineering, and construction. Users can drag and drop model files to view, share, analyze, and comment without opening a desktop design tool. Speckle Intelligence works on published model data and shows the queries, models, and rows behind its answers. Model validation checks published versions against configurable standards and scores each package. Design Review features include a 3D model viewer, issues, markups, and presentations. Listed integrations span ACC, IFC, Revit, Power BI, Rhino, SketchUp, Navisworks, Grasshopper, Blender, Archicad, Civil 3D, AutoCAD, ETABS, Tekla, Vectorworks, Plant 3D, Databricks, and Snowflake. The free Explore plan costs nothing and includes one project, up to three users, 100 versions across a workspace, and 30-day version history. Speckle states that it is SOC 2 Type II certified and aligned with GDPR, and that it does not train AI on customer data. Data is encrypted with TLS in transit and AES-256 at rest; audit logs are marked as coming soon.