Rocket Tools: Aerospace Engineering Intelligence
An aerospace computation engine with 81 validated tools, exposed as a Python library, a CLI, and an MCP server for AI agents. Published on PyPI under Apache 2.0.
Human Engine Labs
The Challenge
Aerospace calculations are scattered across textbooks, MATLAB scripts, and online calculators. Engineers waste time re-deriving basic formulas and verifying unit conversions.
AI agents that assist with engineering tasks lack access to structured, tested tools. They hallucinate formulas, confuse units, and produce unverifiable results.
Design reviews require chaining multiple calculations. Structural loads feed into material selection, which feeds into aerodynamic and trajectory analysis. Siloed tools do not compose.
Our Approach
We built 81 precision tools covering the aerospace stack: structures, aerodynamics and compressible flow, propulsion, orbital mechanics, static stability, ascent and recovery, optimization, and reliability. Every tool is validated against a published reference, with 823 tests and 31 benchmarks.
We exposed the toolset as an MCP (Model Context Protocol) server, so AI agents can call verified engineering functions with structured inputs and outputs. No more hallucinated formulas.
The natural-language router translates plain-english questions into precise tool calls. Ask 'What is the Reynolds number at 5 km and 250 m/s?' and get a verified answer with the full working.
The YAML workflow engine chains tools into multi-step design reviews. A single file can size a vehicle, fly its ascent, and plot the result, then produce a unified report.
Deliverables & Outcomes
What we delivered
- 81 MCP tools across structures, aerodynamics, compressible flow, propulsion, orbital mechanics, static stability, ascent and recovery, optimization, visualization, and reliability
- MCP server for AI agent integration, plus a Python library and a CLI
- Static stability (Barrowman center of pressure and margin) and a full two-body astrodynamics suite: Lambert solver, orbital-element conversion both ways, and Kepler propagation
- Ascent trajectory simulation, vehicle sizing, and parachute recovery sizing
- Motor thrust-curve analysis, plus thermal and pressure-vessel stress for structures
- Design review and FMEA reports for reliability analysis, with provenance and Monte-Carlo uncertainty on every result
- 823 tests, 31 reference benchmarks, published on PyPI (pip install rocket-tools)
Measured results
81
MCP Tools
structures to trajectories
823
Tests
with continuous CI
54 ns
ISA Lookup
cached, per point
Project Gallery
A chemical-equilibrium front end so the propulsion chain closes without an external CEA run, a fuller trajectory model with staging and Mach-dependent drag, and broader structures and astrodynamics coverage. The roadmap is public on GitHub.
See how the protocol works
Each case study demonstrates how Respectful Mediation, Structured Cultivation, and Natural Harmony operate at scale.