reopt Inc. developer resources

What is open

Everything below is public, read-only, and free. No API key, no sign-up, no quota. Call the URLs directly. Nothing here mutates state.

• No authentication — no surface asks for a token or key
• Read-only — every response carries the same content as the public pages
• Courtesy — keep it under one request per second; fetch llms-full.txt once instead of crawling

reopt company site MCP server

The Model Context Protocol endpoint is https://company.reopt.ai/mcp. It speaks Streamable HTTP at revision 2026-07-28 and also accepts handshake-era clients (2025-11-25, 2025-06-18, 2025-03-26). POST only; GET and DELETE return 405.

• get_company_profile — what reopt Inc. is, its mission, operating beliefs, stats, and contact channels
• list_solutions / get_solution — the reopt studio · qa · design · data lineup and per-product detail
• list_principles — the 3O principles (Open · Observe · Optimize) and their public artifacts
• list_pages — every canonical page with its HTML and markdown URL
• get_page_markdown — a page's full body as markdown

curl -s https://company.reopt.ai/mcp \
  -H 'content-type: application/json' \
  -H 'accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2026-07-28' \
  -H 'Mcp-Method: tools/list' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28","io.modelcontextprotocol/clientCapabilities":{}}}}'

Markdown content negotiation

Every page answers `Accept: text/markdown` on the same URL with a markdown representation (acceptmarkdown.com). If you cannot set headers, append `.md` to the canonical URL. Unsupported media types get a 406.

curl -sH 'Accept: text/markdown' https://company.reopt.ai/en/about
curl -s https://company.reopt.ai/en/about.md

OpenAPI document

An OpenAPI 3.1 document lives at https://company.reopt.ai/openapi.json, describing the parameters and status codes of every surface on this page.

Discovery files

Reading one of these first is faster than crawling the site.

• https://company.reopt.ai/llms.txt — when to use this site and how to call it
• https://company.reopt.ai/llms-full.txt — every page, both languages, one document
• https://company.reopt.ai/sitemap.xml · /feed.xml · /atom.xml · /feed.json
• https://company.reopt.ai/.well-known/mcp.json — pointer to the MCP endpoint (convenience, not a standard)

Permission model

These surfaces return the same content as the public web pages, so there is no boundary that would separate one caller from another. That is why no OAuth scope or API-key role exists and why securitySchemes in the OpenAPI document is empty. If an endpoint ever reads non-public data or mutates state, it will arrive with an explicit security scheme, documented here.

Citation and use

When citing this content, attribute the canonical origin (company.reopt.ai). Content terms are on the terms page; questions go to contact@reopt.ai.