I’ve been building Calcurious — a math tool that solves problems step-by-step and generates dynamic visuals (graphs, geometry, symbolic breakdowns) for each step. Each part of the reasoning can be expanded with a “step chat” for deeper explanations. The LLM handles reasoning, but the diagrams + visualization manim engine are fully custom.
I’m looking for feedback from people who care about correctness and clarity. It’s truly inspiring to see the space evolving so quickly! I understand that recent advancements to the Gemini 3 model are a breakthrough in how we will all learn math even better, with its PhD-level reasoning and state-of-the-art scores on challenging math benchmarks like MathArena Apex and AIME 2025.
The focus on enhanced reasoning, multimodal understanding (for interpreting diagrams and video), and agentic capabilities for multi-step planning is a profound validation of the direction Calcurious is heading with its custom visualization and step-chat features. We're all pushing toward a future where complex subjects like math are genuinely accessible to everyone. I'd love to hear your thoughts on how Calcurious can best serve learners and educators. I’m especially interested in the features you think will be essential as these advanced models become more widespread.
I’d appreciate thoughts on: where the reasoning is unclearvisuals that help vs visuals that distractproblem classes where it failsperformance issues, or rough UX edges Features We’d Love to See in Math Tools What would your ideal AI-powered math learning assistant do? Given the immense potential of these new models, here are a few ideas for features we'd love to explore.
Which of these—or others you suggest—would be most impactful for you?
Error Analysis & Correction: A mode that analyzes a user's incorrect step, not just to show the right answer, but to diagnose the specific conceptual misunderstanding (e.g., "You confused the product rule with the chain rule here") and provide targeted remedial practice on just that concept.
Proof Generation/Debugging: For higher-level courses (like Abstract Algebra or Real Analysis), a feature that can help a student construct or debug a mathematical proof with the same step-by-step, explainable detail that Calcurious provides for calculations.
Interactive Simulation Generation: Beyond static graphs, what if the tool could generate a live, interactive simulation or mini-game based on the problem (e.g., a projectile motion problem generates a small game where you adjust launch angles)?
"What If" Scenarios: The ability to instantly tweak variables in the problem and see how the final solution, the steps, and the visualizations dynamically change. (e.g., "What if the spring constant was 2k instead of k?")
Hey HN,
I’ve been building Calcurious — a math tool that solves problems step-by-step and generates dynamic visuals (graphs, geometry, symbolic breakdowns) for each step. Each part of the reasoning can be expanded with a “step chat” for deeper explanations. The LLM handles reasoning, but the diagrams + visualization manim engine are fully custom.
I’m looking for feedback from people who care about correctness and clarity. It’s truly inspiring to see the space evolving so quickly! I understand that recent advancements to the Gemini 3 model are a breakthrough in how we will all learn math even better, with its PhD-level reasoning and state-of-the-art scores on challenging math benchmarks like MathArena Apex and AIME 2025.
The focus on enhanced reasoning, multimodal understanding (for interpreting diagrams and video), and agentic capabilities for multi-step planning is a profound validation of the direction Calcurious is heading with its custom visualization and step-chat features. We're all pushing toward a future where complex subjects like math are genuinely accessible to everyone. I'd love to hear your thoughts on how Calcurious can best serve learners and educators. I’m especially interested in the features you think will be essential as these advanced models become more widespread.
I’d appreciate thoughts on: where the reasoning is unclearvisuals that help vs visuals that distractproblem classes where it failsperformance issues, or rough UX edges Features We’d Love to See in Math Tools What would your ideal AI-powered math learning assistant do? Given the immense potential of these new models, here are a few ideas for features we'd love to explore.
Which of these—or others you suggest—would be most impactful for you?
Error Analysis & Correction: A mode that analyzes a user's incorrect step, not just to show the right answer, but to diagnose the specific conceptual misunderstanding (e.g., "You confused the product rule with the chain rule here") and provide targeted remedial practice on just that concept.
Proof Generation/Debugging: For higher-level courses (like Abstract Algebra or Real Analysis), a feature that can help a student construct or debug a mathematical proof with the same step-by-step, explainable detail that Calcurious provides for calculations.
Interactive Simulation Generation: Beyond static graphs, what if the tool could generate a live, interactive simulation or mini-game based on the problem (e.g., a projectile motion problem generates a small game where you adjust launch angles)?
"What If" Scenarios: The ability to instantly tweak variables in the problem and see how the final solution, the steps, and the visualizations dynamically change. (e.g., "What if the spring constant was 2k instead of k?")
Link again for convenience: https://beta.calcurious.ai/
Thanks for checking it out — all critiques are welcome.