Rocket Tools 0.4.0: from calculators to a vehicle that flies
Most aerospace calculators answer one question at a time. What is the drag at this speed? What is the deflection under this load? Useful, but a real design is a chain. The engine sizes the tank, the tank sizes the vehicle, the vehicle flies a trajectory, and the trajectory decides whether any of it was a good idea.
Rocket Tools 0.4.0 closes that loop. It is now on PyPI (pip install rocket-tools), with 68 validated tools you can call from Python, a CLI, or an MCP server.
What is new
The headline is a real ascent simulator. simulate_ascent integrates a point-mass vehicle through the full ISA atmosphere with a fixed-step RK4 method, accounting for thrust, drag, and altitude-varying gravity. It reports burnout and apogee, peak dynamic pressure, and the g-load along the way. The integrator is pinned to the closed-form vacuum trajectory, so its own error is checked against theory before any atmosphere or gravity model is added.
Sizing comes with it. size_vehicle solves the rocket equation for gross mass and chains the existing thrust-to-weight and tank-sizing tools, so a delta-v budget turns into a full mass breakdown in one call.
We also added optimization. optimize_staging finds the payload-maximizing delta-v split across stages with a Lagrange solver, validated against an independent brute-force search. optimize_design is the general version: give it any tool, any input variable, and an output to maximize or minimize, and it runs a golden-section search over the result.
Two more areas round out the release. The visualization tools return either the plotted data or a native image, so an agent can show a shear-moment diagram or a trajectory plot inline. The reliability tools roll margins of safety into a pass/fail design review and rank failure modes by risk priority number.
Why it matters for agents
Every number Rocket Tools produces is validated against a published reference, and cite_tool returns the source, formula, and assumptions behind it. An AI agent can now do more than recall a formula. It can size a vehicle, fly its ascent, plot the result, and show its working, without leaving the conversation.
Getting started
pip install rocket-tools
The source, the full changelog, and the tool reference are on GitHub. If you build something with it, we would like to see it.