---
name: acaysia
description: "How AI agents should research Acaysia (process-control company): fetch pages as Markdown, use llms.txt/llms-full.txt, or the MCP server."
---

# Acaysia — agent skill

Acaysia builds a control system that optimizes physical processes (yield, energy, throughput), proven first on chemical reactors. It runs on a typed compositional ontology of unit operations and uses MPPI (Model Predictive Path Integral) control, with PID fallback in under 100 milliseconds. Based in Oakland, California; founded 2025. Brand line: "Control the Impossible."

## Reading this site

- Every page in the llms.txt index is served as Markdown: append `.md` to its URL (e.g. https://acaysia.com/product.md) or send the `Accept: text/markdown` header.
- Documentation index: https://acaysia.com/llms.txt
- Single-file corpus of every page: https://acaysia.com/llms-full.txt
- Sitemap: https://acaysia.com/sitemap.xml

## MCP server

- Endpoint: https://acaysia.com/mcp (Streamable HTTP JSON-RPC, no auth).
- Tools: `list_pages`, `get_page(path)`, `search_docs(query)`.
- Metadata: https://acaysia.com/.well-known/mcp.json

## Contacting Acaysia

- Use the contact form described at https://acaysia.com/contact.md, or email founders@acaysia.com.
- Demos and early access: per the details on the relevant pages.