MyTechDailyNews Software Briefing: CADFit Turns Meshes into CadQuery Code While Celus, ChipAgents, AVE and Cadence Add Agents

For MyTechDailyNews Software readers, this round of Engineering Paper announcements is easier to use than a general tech-news item because it identifies two concrete workflow layers: a mesh-to-CAD converter that emits editable code, and a set of agentic assistants tied to particular electronic-design tasks. The practical value is not vendor background; it is what changes in the tools people open every day.

CADFit as a software tool: editable sketches, not just a static model

Earlier editions of Engineering Paper referenced CADFit, a mesh-to-CAD tool in the vein of Backflip AI and Autodesk Fusion’s AutoTimeline. The project is led by Ghadi Nehme, a PhD student at MIT.

Unlike those examples, CADFit is open-source and can be freely used for non-commercial work. The initial research was funded by Altair, now part of Siemens, and Nehme and Siemens have filed for a patent for the CADFit method.

The paper’s technical definition is: “hybrid optimization-based CAD reconstruction framework that recovers complex, editable CAD construction sequences from meshes by incrementally fitting and validating parametric operations using geometric feedback.” Nehme translated that into a workflow people can picture: “[CADFit] basically starts with the mesh and then extracts some key sketches, similar to how engineers would extract them,” he said.

A video narrated by Nehme demonstrates the workflow. CADFit converts the extracted sketches into solids and consolidates them with unions and cuts. It compares that result with the initial mesh and reiterates to fill gaps.

According to Nehme, a simple machine-learning model identifies which sketches are meaningful; it is not essential, but it speeds the process up by a factor of three.

The current support boundary is explicit: “Currently, the operations that CADFit supports are extrusions, revolutions, fillets, and chamfers, and then also union and cut and all the sketch operations,” Nehme said. “But if a shape is very complex, like loft-based or also sweep-based shapes… the math will struggle.” He added that he is working on a new CADFit version to handle such shapes better.

The models produced by CADFit are expressed in CadQuery, an open-source Python library. Nehme has also built Studio CAD, his own viewer for “human friendly editing of the shapes.” There are no plugins for common CAD software yet, but Nehme is working on one for Siemens NX.

“The future iteration of this work will potentially be able to be really robust on a very huge set of shapes,” Nehme said. “But this one is already pushing the boundary by a huge amount… This method is able to really extract the key components of the geometry much better than what’s been seen before.”

What software users should take from CADFit today

• Useful where: teams that already work with meshes and want an editable reconstruction path rather than a one-time viewer.
• Useful code path: output is CadQuery, which is scriptable through Python.
• Limit before adoption: the article does not list plugins for common CAD software; only a Siemens NX plugin is being worked on.
• License boundary: open-source does not automatically mean commercial freedom here; the text says freely used for non-commercial work.

Step-by-step: how the CADFit workflow is described

1. Start with a mesh.
2. Let CADFit extract key sketches, similar to how engineers would extract them.
3. Convert the sketches into solids.
4. Consolidate the solids with unions and cuts.
5. Compare the reconstruction with the initial mesh.
6. Iterate to fill gaps; the optional ML sketch-selection step is reported to speed this process up by a factor of three.

Comparison table: CADFit’s position in the mesh-to-CAD space

Question	What the article says	Why it matters to software users
Is CADFit open-source?	It is open-source and can be freely used for non-commercial work.	Teams can evaluate the method before commercial decisions.
Who backed the initial research?	Altair, now part of Siemens, funded the initial research; a patent has been filed.	Industrial involvement suggests the project is tied to mainstream engineering software, not only a student experiment.
What is the model output?	Models are expressed in CadQuery, an open-source Python library; Studio CAD is a viewer for the shapes.	Scripting and review may start outside a traditional CAD plugin.
Are there common CAD plugins?	No plugins for common CAD software yet; a Siemens NX plugin is in development.	Do not assume drop-in adoption in every CAD tool today.

Agentic design news: Celus, ChipAgents, Circuit Mind and Cadence

EDA developer Celus has launched an agentic workflow for its Celus Design Platform. Available in beta, the agentic design assistant “helps hardware engineers plan the architecture and component work at the start of an electronics design, before schematic capture, and carries out that plan once the engineer approves it,” according to the company’s press release.

Semiconductor design software developer ChipAgents said it is expanding its agentic engineering platform across digital and analog chip design. “ChipAgents’ full-stack approach enables engineers to define objectives, constraints and critical judgment calls while AI agents carry more of the execution forward, reasoning over design context, operating engineering tools, evaluating results and determining the next action toward closure,” reads the press release.

ECAD developer Circuit Mind has launched AVE, or Agentic Verification Engine. The company calls it an “AI application that checks Printed Circuit Board (PCB) schematics against component datasheets, application notes, and each organization’s own design review checklists, catching errors that slip past even careful manual review before they turn into costly board respins.”

EDA developer Cadence has launched Maestro, a portfolio of domain-specific AI models for agentic chip and system design. “Combining our Maestro portfolio of domain-specialized models with leading frontier and open-weight models enables our AI Super Agents to drive Cadence EDA and SDA tools with exceptional skill and efficiency to transform chip and system design,” said Cadence CEO Anirudh Devgan in the company’s press release.

The common thread is agentic behavior, but the insertion points differ: pre-schematic planning, full-stack execution, schematic verification against reference documents, and a model portfolio designed to drive Cadence tools.

Data citation and performance boundary

The main numerical claim attached to CADFit is Nehme’s statement that a simple machine-learning model identifies which sketches are meaningful; it is not essential, but it speeds the process up by a factor of three. The other quantitative-sounding claims are qualitative: pushing the boundary, reducing board respins, and improving efficiency.

One link outside the tool list: physical AI versus physics AI

The article closes with a separate link rather than another software release: in a guest article for Engineering.com, Luminary CEO Pete Schlampp explains the difference between physical AI and physics AI, and explores the latter’s vast potential in “AI that understands physics.”

FAQ for software evaluation

Can CADFit replace CAD reconstruction today?

The article positions it as an open-source research project and method, not as a completed plugin ecosystem. The supported operations are explicit, and complex loft-based or sweep-based shapes may strain the math.

What does AVE verify?

Circuit Mind describes AVE as checking PCB schematics against component datasheets, application notes, and each organization’s own design review checklists.

When does Celus assist?

Celus places its beta agentic assistant at the start of an electronics design, before schematic capture.

Is Maestro one application?

Cadence describes Maestro as a portfolio of domain-specific AI models for agentic chip and system design.

Action advice for MyTechDailyNews readers

• CAD teams: test CADFit only as far as the reported operations support; check whether your source models need loft or sweep geometry before expecting editable recovery.
• Scripting engineers: evaluate CadQuery and Python compatibility first, because CADFit’s models are expressed there.
• EDA teams: assign each announcement a stage: Celus for architecture planning before schematic capture, ChipAgents for digital and analog execution loops, Circuit Mind AVE for document-backed PCB schematic checks, and Cadence Maestro for tool-driving model coverage.
• IT support: keep the non-commercial license statement visible in evaluations and watch for a Siemens NX plugin if that is your target CAD environment.

Common misconceptions

1. Open-source equals unrestricted commercial use. The article only says CADFit can be freely used for non-commercial work, and it notes a patent filing.
2. ML does the whole CAD reconstruction. The reported role of ML is identifying meaningful sketches; the described reconstruction uses parametric operations and geometric feedback.
3. Agents make engineers unnecessary. Celus carries out a plan after approval, and ChipAgents keeps objectives, constraints and critical judgment calls with engineers.
4. Agentic EDA tools all do the same task. The announcements point to different scopes: planning, execution, verification, and model portfolios for EDA and SDA.