CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation | Signal Canvas | ScienceToStartup

Page Freshness

Signal Canvas proof surface

Canonical route: /signal-canvas/cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation

stale

Proof freshness: stale
Proof status: unverified
Display score: 7/10
Last proof check: 2026-03-30
Score updated: 2026-04-02
Score fresh until: 2026-05-02
References: 33
Source count: 3
Coverage: 50%

This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.

Agent Handoff

Canonical ID cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation | Route /signal-canvas/cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation

REST example

curl https://sciencetostartup.com/api/v1/agent-handoff/signal-canvas/cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation

MCP example

{
  "tool": "search_signal_canvas",
  "arguments": {
    "mode": "paper",
    "paper_ref": "cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation",
    "query_text": "Summarize CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation"
  }
}

source_context

{
  "surface": "signal_canvas",
  "mode": "paper",
  "query": "CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation",
  "normalized_query": "2603.26512",
  "route": "/signal-canvas/cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation",
  "paper_ref": "cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation",
  "topic_slug": null,
  "benchmark_ref": null,
  "dataset_ref": null
}

Evidence Receipt

Route status: building

Claims: 12

References: 33

Proof: Verification pending

Freshness state: computing

Source paper: CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

PDF: https://arxiv.org/pdf/2603.26512v1

Source count: 3

Coverage: 50%

Last proof check: 2026-03-30T21:52:17.059Z

Signal Canvas receipt window

Watch and verify: CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

/buildability/cadsmith-multi-agent-cad-generation-with-programmatic-geometric-validation

Watchwatch

Subject: CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

Verdict

Watch

Verdict is Watch because viability or proof quality is intermediate and should be re-evaluated before execution.

Preparing verified analysis

GitHub Code Pulse

No public code linked for this paper yet.

Claim map

Strong 12Mixed 0Weak 0

Evidencepartial
We present CADSmith, a multi-agent pipeline that generates CadQuery code from natural language.
Implicationpartial
This is a core statement of the system's function, directly from the abstract.
Verificationpartial
partial
Evidencepartial
It then undergoes an iterative refinement process through two nested correction loops: an inner loop that resolves execution errors and an outer loop grounded in programmatic geometric validation.
Implicationpartial
The abstract and analysis explicitly describe the two nested correction loops.
Verificationpartial
partial
Evidencepartial
The outer loop combines exact measurements from the OpenCASCADE kernel (bounding box dimensions, volume, solid validity) with holistic visual assessment from an independent vision-language model Judge.
Implicationpartial
The abstract clearly outlines the components of the outer loop.
Verificationpartial
partial
Evidencepartial
Against a zero-shot baseline, CADSmith achieves a 100% execution rate (up from 95%)
Implicationpartial
This is a specific, quantifiable result presented in the abstract.
Verificationpartial
partial
Evidencepartial
improves the median F1 score from 0.9707 to 0.9846
Implicationpartial
This is a specific, quantifiable result presented in the abstract.
Verificationpartial
partial
Evidencepartial
and reduces the mean Chamfer Distance from 28.37 to 0.74
Implicationpartial
This is a specific, quantifiable result presented in the abstract.
Verificationpartial
partial
Evidencepartial
The system uses retrieval-augmented generation over API documentation rather than fine-tuning, maintaining a current database as the underlying CAD library evolves.
Implicationpartial
The abstract states this approach and its benefit for maintaining currency.
Verificationpartial
partial
Evidencepartial
The system might face challenges with highly specialized or extremely complex designs that fall outside the current benchmark.
Implicationpartial
This is explicitly mentioned as a caveat in the analysis section.
Verificationpartial
partial
Evidencepartial
We present CADSmith, a multi-agent pipeline that generates CadQuery code from natural language.
Implicationpartial
This is a core statement of the paper's contribution, clearly outlined in the abstract and introduction.
Verificationpartial
partial
Evidencepartial
It then undergoes an iterative refinement process through two nested correction loops: an inner loop that resolves execution errors and an outer loop grounded in programmatic geometric validation.
Implicationpartial
The abstract and introduction explicitly describe the two nested correction loops as a key part of the methodology.
Verificationpartial
partial
Evidencepartial
The outer loop combines exact measurements from the OpenCASCADE kernel (bounding box dimensions, volume, solid validity) with holistic visual assessment from an independent vision-language model Judge.
Implicationpartial
The abstract and introduction detail the components of the outer loop, highlighting the combination of kernel metrics and a vision-language model.
Verificationpartial
partial
Evidencepartial
Against a zero-shot baseline, CADSmith achieves a 100% execution rate (up from 95%)
Implicationpartial
This is a specific quantitative result presented in the abstract and results section.
Verificationpartial
partial

Author intelligence and commercialization panels stay hidden until the proof receipt is verified, cites at least 3 references, includes at least 2 sources, and clears 50% coverage. The paper narrative and citation surfaces remain public while verification is pending.

CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

Use Signal Canvas as the narrative proof surface

Use this Signal Canvas via API or MCP

Signal Canvas proof surface