compounders.ai

compounders.ai for AI agents

These pages are plain HTML that reads without JavaScript, and each has a Markdown copy. So does each company's overview page; the rest of the terminal runs in the browser, at addresses that follow a fixed grammar an agent can build.

Updated · Ruleset 1.4.0 · Markdown version

What can an agent read without running JavaScript?

The rest of the terminal (the screener, the company tabs, Universe, Compare, Plans) is rendered in the browser and needs JavaScript; an agent that does not run it should read these docs and the overview pages, then link the tab it means.

How are company page addresses built?

Every company page is addressed by its listing key, <exchange>-<ticker>, and every address ends in a slash:

PageAddress
Overview/company/<key>/
Income statement/company/<key>/financials/income-statement/annual/
Balance sheet/company/<key>/financials/balance-sheet/annual/
Cash-flow statement/company/<key>/financials/cash-flow-statement/annual/
Normalized figures and returns/company/<key>/financials/normalized-and-returns/annual/
Normalization ledger/company/<key>/normalization/
Growth/company/<key>/growth/
Company KPIs/company/<key>/kpis/
Returns/company/<key>/returns/
Sources/company/<key>/sources/

Examples: /company/TSX-CSU/ (Constellation Software), /company/OM-LIFCO-B/ (Lifco), /company/NasdaqGS-ROP/ (Roper Technologies). A bare ticker or another spelling of the key redirects to the listing key. Every key is on the Data coverage page.

Which exchange codes do listing keys use?

CodeExchange
TSXToronto Stock Exchange
TSXVTSX Venture Exchange
OMNasdaq Stockholm
NasdaqGSNasdaq Global Select Market
LSELondon Stock Exchange
AIMLondon Stock Exchange, AIM
NYSENew York Stock Exchange
XAMSEuronext Amsterdam
XCSENasdaq Copenhagen
XPAREuronext Paris
XOSLOslo Børs
XMILEuronext Milan

Is there an API or an MCP server?

Not at present. compounders.ai does not offer a public API or MCP server; its figures are read on the site, where each opens its citation.

Which crawlers may read the site?

All of them. /robots.txt allows search engines and AI crawlers, including Googlebot (whose rules Brave Search's crawler also follows), Bingbot, Applebot, OAI-SearchBot, GPTBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot and Perplexity-User, and asks crawlers to skip only the sign-in pages.

How should an agent quote a figure?

With the measure's exact name, the fiscal year, the currency, the ruleset version and a link to the company page; a withheld figure is reported as withheld, never filled in. The Citation guidelines have the format.