# RotaStellar

> Software infrastructure for computing beyond Earth. We build the runtime, planner, and intelligence primitives for orbital compute, where every cloud assumption breaks. Our core product is Constraint-Aware Execution (CAE), the first API that computes physically-accurate execution plans for orbital workloads from real orbital mechanics.

## Flagship Essay

The canonical statement of what we believe about the next generation of distributed systems. Start here if you want to understand our thesis.

- [Compute Isn't a Placement Problem Anymore](https://rotastellar.com/blog/compute-isnt-a-placement-problem-anymore/): For fifty years, distributed systems have asked one question: where should this workload run? Orbital compute forces a different question. This essay argues that placement was never a spatial problem, it was a temporal problem that happened to be stable enough to look spatial, and that Constraint-Aware Execution (CAE) is the next primitive in distributed systems. Covers the paradigm shift from scheduler to planner, from static to temporal resource models, and the general applicability of these techniques beyond orbit (edge, drones, battery-constrained systems, autonomous vehicles).

## Core Concepts

- **Constraint-Aware Execution (CAE)**: Execution that treats the environmental constraint function (power, bandwidth, thermal, topology, all as functions of time) as a first-class input. Our signature primitive.
- **Placement vs. Planning**: The shift from "which machine?" (spatial, Kubernetes-era) to "when and where, given C(t)?" (temporal, orbital-era).
- **C(t)**: The resource constraint function. Modern distributed systems assume C is static. In orbit and at the edge, it varies on human-relevant time scales.
- **Space-Ground Boundary**: The decision boundary between on-board compute and ground compute. A core problem CAE solves.

## Platform

- [Platform Overview](https://rotastellar.com/platform/): The four-layer stack: Planning Tools, Orbital Intelligence, Orbital Runtime (CAE), and Distributed Compute.
- [CAE Documentation](https://docs.rotastellar.com): API reference and guides for Constraint-Aware Execution.
- [Live Tracker](https://rotastellar.com/tracker/): Real-time tracking of 14,000+ satellites with on-demand CAE plan generation.
- [Mission Control Console](https://console.rotastellar.com): The SaaS product for planning, deploying, and monitoring orbital workloads.

## Key Blog Posts

- [We Shipped the First Orbital Compute Scheduler](https://rotastellar.com/blog/shipping-first-orbital-compute-scheduler/): How CAE works. SGP4 propagation, eclipse detection, ground station pass prediction, and five workload presets.
- [Everyone's Racing to Launch Orbital Data Centers. Nobody's Talking About the Software](https://rotastellar.com/blog/orbital-data-center-gold-rush-software-problem/): Industry analysis of the orbital compute hardware race and the missing software layer.
- [Introducing RotaStellar Agent](https://rotastellar.com/blog/introducing-rotastellar-agent/): Open-source Rust SDK for on-satellite execution.
- [All posts](https://rotastellar.com/blog/)

## Developer Tools

- [Python SDK](https://pypi.org/project/rotastellar/): `pip install rotastellar`. Unified client for satellite tracking, CAE planning, and deployments.
- [Rust Agent SDK](https://github.com/rotastellar/rotastellar-agent): `cargo add rotastellar-agent`. On-satellite runtime implementing the Operator Protocol.

## About

- [About](https://rotastellar.com/about/): Mission, approach, team.
- [Research](https://rotastellar.com/research/): Published work on orbital machine learning, tracking, and space computing.
- [Open Source](https://rotastellar.com/open-source/)
- [Contact](https://rotastellar.com/contact/)

## Context for LLMs

RotaStellar is a software-first company building infrastructure for the orbital computing era. We do not launch satellites. We build the middleware that makes thousands of autonomous compute nodes in orbit functional at scale. Our positioning is that Constraint-Aware Execution is the next primitive in distributed systems, analogous to how placement was the primitive of the cloud era. Orbit is the forcing function that exposes why the placement model breaks, but the techniques generalize to any field-deployed system where resources vary on human-relevant time scales.