Investigate a problem with the assumptions in view
The assistant shows the data used, the simplifications, and what is still missing before proposing a path.
AI for engineering and calculation
In researchEffect.LM
A technical assistant for investigating engineering problems without hiding the assumptions.
Not available yet: features, interface, and requirements may change. Availability will be announced on this page.
Context providedProject fileTechnical referenceExternal servicesDisclosed
Question
What assumptions are being used in this calculation?
- AssumptionsTo check
- Units and dimensionsTo check
- MethodRecorded
- SourcesLinked
- LimitsExplicit
Human review pendingRejectAccept
What you can do with it.
Check units and conversions
Units and dimensions appear next to every result, with inconsistencies flagged before you accept.
Repeat and review the calculation
Inputs, method, and sources are recorded, so the calculation can be repeated and checked by someone else.
How it works.
Frame the question
Define the problem, the known data, the units, and what still needs investigating.
Examine the reasoning
Explore assumptions, methods, and references, with the points that need checking marked.
Verify and decide
Review sources, calculations, and risks before accepting or rejecting a suggestion.
Choose the right version for your machine.
Effect.LM will come in three versions; what changes is the computer each one calls for. Choose based on the memory and graphics card you have. All three are in research, and the requirements may change.
Which version fits your computer?
Recommended for your computerEffect.LM Raptor
This selection happens only in your browser. The requirements are recommendations under evaluation, not guarantees.
Effect.LM Raptor Pro
The most advanced
For workstations, the high-performance computers used in engineering. Intended for complex problems, multi-step reasoning, and agents, where the AI organizes the work and uses calculation tools.
- Recommended computer
- Graphics card (GPU) with 24 GB of memory, or 32 GB of RAM
- Base technology
- Open models from the IBM Granite family
Effect.LM Raptor
The most balanced
Not sure? Start here: between the capability of Raptor Pro and the lightness of Flash, designed for everyday laptops and desktops.
- Recommended computer
- 16 GB of RAM; graphics card optional
- Base technology
- Open models from the IBM Granite family
Effect.LM Flash
The lightest and fastest
For entry-level computers, even without a graphics card: everyday calculations, unit conversions, and quick checks.
- Recommended computer
- 8 GB of RAM, even without a graphics card (CPU only)
- Base technology
- Compact reasoning model
What every version includes.
- They run on your computer, with a clear notice whenever something needs the internet or an external service.
- The AI plans and explains the reasoning; the math is handled by calculation tools that always give the same result for the same inputs.
- Assumptions, units, and sources in plain view, so you can check them.
- No decision without your review.
Where the product stands today.
- Next milestone
- Validate use cases and verification criteria with professionals.
- Where it runs
- On your computer, in three versions: Raptor Pro, Raptor, and Flash.
- Initial focus
- Biomedical and computational engineering.
- In the Effect family
- Integration with Feito 4D Studio, on the roadmap.
Frequently asked questions.
Can I use Effect.LM yet?
Not yet. Effect.LM is in research, and the start of testing will be announced on this page.
Does the AI do the math on its own?
No. The AI plans and explains the reasoning; the math is handled by calculation tools, with units checked. Human review remains mandatory.
Does my data leave my computer?
The plan is to do the processing on your computer and show exactly what was shared with the AI. If any step depends on the internet or an external service, you will be told.
Want to follow the next steps?
Follow the progress of Effect.LM and its first tests.