Skip To Content
hamburger
hamburger
Cadmould API Toolkit

The injection 
molding simulation 
API your agent loves

An API for plastic injection molding simulation, built for automated workflows and AI agents from day one. Up to 1,000x faster than classical solvers. Built by engineers, for engineers who build.
Cadmould API Toolkit Ilustration
  • Up to 1,000x faster

    than classical solvers for filling analysis

  • 2-5% error

    on most filling cases vs. validated numerical simulation

  • One API

    Neural and classical engines, one parameter switch

  • Run in parallel

    Sweep hundreds of variants simultaneously: Your agent fans out

Use Cases

What you can build with the Cadmould API Toolkit

From a quick DFM check to a 500-run parameter sweep - all from your code.

AI agent integration

Give your agent a physics engine. Cadmould API Toolkit handles the simulation, the agent processes the results, 
you pick the winner.

Automated 
DfM pipeline

Wire a fill check into your geometry pipeline. Auto-place a gate, run the fill, and get a red/amber/green signal on short shots, weld lines, air traps, and clamp force - before you sit down to review the design.

Robustness 
studies

If your material properties vary by a few percent, how do your outcomes vary? Run the Monte Carlo study you have never had time for. Fast physics makes it practical at the design stage, not just in final validation.

Internal 
simulation tooling

Wrap the API into any internal tool your team already uses. They run simulations without touching the complexity. Build once. Run by anyone.

Gate placement sweeps

Sweep dozens of gate locations, score each fill by pressure and defect rate, and let your agent pick the winner. What used to be 10 designs overnight becomes 10,000.

Get Started

Built by engineers, for engineers who build.

Wave 1 is underway. A small group of engineers are already running simulations, shaping the API, and getting direct access to the team who built it. Apply to join the next wave.

Request early access
  • Get the starter kit with working simulation examples
  • Direct access to the SIMCON engineering team
  • A say in how the API evolve

Get Started

Five steps. One import.

  1. 01

    Upload your geometry

    STL or Cadmould RM1. Registered as a platform reference. Upload once, reuse across unlimited simulations.

  2. 02

    Select a material

    Search by trade name, material family, or fiber content - partial names work. Pick yours from the results. Cloud-based and always current.

  3. 03

    Set your process

    Melt temperature, injection rate, mold temperature, gate positions. 
All inline with the simulation call.

  4. 04

    Run a simulation

    engine='neural' for seconds. engine='numerical' for classical precision. The SDK handles the job lifecycle.

  5. 05

    Get results

    Flow front, pressure, temperature, weld lines, clamping force.
JSON or HDF5. Post-process locally or pipe to the next step.

Screenshot of cadmould SDK code and a colorized 3D simulation result

Powered With

The dual engine: explore at speed, verify with confidence.

Agents are good at planning an coordination. But can they predict non-Newtonian flow? Just give your agent access to our dual engine and they can orchestrate and run your experiments with speed and confidence.

The AI Solver handles hundreds of filling variants in the time a classical solver takes one. When you've found the candidates worth validating, one parameter switch moves to the classical numerical solver for full-fidelity validation - same API, same geometry, same material reference.

hundreds of filling variants at once

The AI Solver

  • engine="neural"
  • Up to 1,000x faster than classical solvers
  • Full 3D fill results in seconds
  • High-fidelity (<5% error on most filling cases)
  • Built for high-volume sweeps and automated workflows

1,000×

faster than FEA

seconds

per simulation

<5%

typical error

full-fidelity verification

Numerical Solver

  • engine="numerical"
  • Established, validated physics
  • Consistent with the production simulation tools you already know
  • Same API, same geometry, same material reference as the AI Solver

~3-8 minutes

per simulation

FEA-grade

accuracy

< 0.5%

error (est.)

AI Solver scope today

Filling phase: flow front, pressure, weld lines, air traps, clamp force. Cooling, warpage, and fiber-reinforced grades are not yet modeled — switch to classical for full-fidelity validation.

See our documentation
Request early access

Wave 1 is underway. Sign up to join the next wave.

Start building today.

Right now, a handful of engineers are running injection molding simulations in seconds. They have a direct line to the product team and workflows nobody else has built yet. Sign up to join the next wave.

What happens when you apply

  • We review your application and confirm fit
  • You receive the starter kit to explore the API
  • If it's a match, we set up your access and onboard you directly

Firefly Isometric 3D illustration of a plastic injection molding fill simulation.

Connects to everything you already use.

Open formats throughout. Your tools can read every input and every output. No proprietary middleware required.

  • Python
  • PyPI
  • Claude Code
  • n8n
  • HDF5
  • JSON
  • pandas
  • STL
  • Git
  • Jupyter

Questions from builders

What geometries does the API accept?

You upload the part as an STL mesh or a Cadmould RM1 project file. (If you mesh the part with the SDK, you can upload the resulting .cfex directly.) Geometry is registered once on the platform and then referenced by id, so a large multi-run study uploads the part a single time instead of resending it with every simulation.
Still have questions?
We have covered a few here. Explore our full FAQ section for detailed insights and answers to your questions.