Equity Rank for AI assistants and developers

Equity Rank recomputes 8–15 valuation models for roughly 3,000+ U.S.-listed securities after every market close. Machines can read the result two ways: an MCP server that AI assistants call at answer time, and public JSON endpoints for anything else. Neither needs an account or a key.

MCP server

The Model Context Protocol is the open standard AI assistants use to call external tools. Equity Rank's server speaks the Streamable HTTP transport, stateless, with no authentication:

https://api.equity-rank.com/mcp

Once connected, the assistant can look up a security and quote the live figure with a link to its Equity Rank page. Four read-only tools:

Tool	What it returns
`search`	Resolve a ticker or company name to securities in the coverage universe (up to 10 matches, each with its page URL).
`fetch`	A readable valuation summary for one ticker — price, model fair value estimate, margin-of-safety differential, method agreement, Fundamental Strength, next earnings, last-updated stamp — with structured metadata and the source URL.
`get_stock_valuation`	Structured public payload for one ticker (identical to GET /api/stocks/{TICKER}/public).
`list_sector_stocks`	The public sector index: top scored securities in a sector, ranked by Equity Rank Score, public-safe fields only.

search and fetch follow the connector contract ChatGPT requires, so the same server works in every client. Rate limit: 120 requests per minute per IP.

Claude (claude.ai)

Settings → Connectors → Add custom connector. Name it Equity Rank, paste the URL above, leave the OAuth fields empty. Then enable it in a chat from the tools menu.

Claude Code

claude mcp add --transport http equity-rank https://api.equity-rank.com/mcp

ChatGPT

Settings → Connectors → Advanced → enable Developer mode → Create. Paste the URL, choose No authentication, and save. Select the connector in a new chat.

Cursor and other MCP clients

{
  "mcpServers": {
    "equity-rank": { "url": "https://api.equity-rank.com/mcp" }
  }
}

VS Code (Copilot agent mode)

Add to .vscode/mcp.json in the workspace, or to the user settings under the same key:

{
  "servers": {
    "equity-rank": { "type": "http", "url": "https://api.equity-rank.com/mcp" }
  }
}

Windsurf

In ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "equity-rank": { "serverUrl": "https://api.equity-rank.com/mcp" }
  }
}

Codex CLI

In ~/.codex/config.toml:

[mcp_servers.equity-rank]
url = "https://api.equity-rank.com/mcp"

Gemini CLI

In ~/.gemini/settings.json:

{
  "mcpServers": {
    "equity-rank": { "httpUrl": "https://api.equity-rank.com/mcp" }
  }
}

Clients that only launch local (stdio) servers

Zed and older clients start a local process rather than opening a URL. Any stdio-to-HTTP relay works; the community mcp-remote package needs no configuration:

{
  "mcpServers": {
    "equity-rank": { "command": "npx", "args": ["-y", "mcp-remote", "https://api.equity-rank.com/mcp"] }
  }
}

Claude Desktop one-click bundle

Where custom connectors are not available, download equity-rank.mcpb and open it with Claude Desktop. The bundle holds a dependency-free local bridge that relays to the URL above; nothing is stored and no key is needed.

ChatGPT without Developer mode

Search the GPT Store for "Equity Rank". Its Actions read the public endpoints through the OpenAPI document below, which any agent framework can also import:

https://api.equity-rank.com/openapi-public.json

Public JSON endpoints

One security — the payload every MCP tool builds on:

GET https://api.equity-rank.com/api/stocks/{TICKER}/public

Fields include company_name, sector, price, consensus_fair_value (the blended model estimate), mos_pct (margin-of-safety differential, %), method_agreement_above / _below / _total, fundamental_strength_score / _grade, next_earnings_date / _eps_estimate, and last_updated. A ticker outside the coverage universe returns HTTP 404 — there are no placeholder rows.

A sector's top scored securities, public-safe fields only:

GET https://api.equity-rank.com/api/screener/sector/{sector}

Sectors: technology, healthcare, financials, consumer discretionary, consumer staples, industrials, energy, materials, real estate, utilities, communication services. Every ticker in coverage is listed at GET https://api.equity-rank.com/api/stocks-index.

Markdown pages

Every stock page has a markdown form, the least-noise version to quote from, with each figure's qualifier inline:

GET https://www.equity-rank.com/stocks/{TICKER}.md

The same document is returned for https://www.equity-rank.com/stocks/{TICKER} when the request sends Accept: text/markdown. The reference pages are bundled into one file for single-fetch readers at /llms-full.txt.

Bulk dataset

The whole public table as one file, one row per covered security, rebuilt after every market close. Same fields as the per-security endpoint, minus the live quote: the price column is the close the batch valued against, and the column names carry the qualifiers.

GET https://api.equity-rank.com/api/datasets/valuations-latest.csv
GET https://api.equity-rank.com/api/datasets/valuations-latest.json

Columns: ticker, company_name, sector, price_at_last_recompute, model_fair_value_estimate, margin_of_safety_differential_pct, models_above_price / models_below_price / models_total, fundamental_strength_score / _grade, equity_rank_score, market_cap, last_recomputed. The JSON form carries a meta block with the column dictionary, the batch timestamp and the terms; GET https://api.equity-rank.com/api/datasets returns that block alone. Responses carry an ETag and Last-Modified, so a daily fetch with If-None-Match costs nothing when the batch has not run.

Free to use with attribution to Equity Rank and a link back, under the terms of service. Every value is a model estimate under stated assumptions; the citation rules below apply to the file exactly as they do to a page.

How to quote a figure

Equity Rank's public data may be quoted with attribution to "Equity Rank" and a link to the source URL. These qualifiers are part of the claim and travel with any number:

Fair value is a model estimate under stated assumptions — not a target price, not a prediction, not a recommendation.
Margin of safety is the percentage differential between the model estimate and the current price. It describes model output; it does not predict returns.
Method agreement ("N of M models place fair value above the current price") describes dispersion among models, not confidence in an outcome.
Carry the date. Figures are recomputed nightly; every payload includes last_updated.

The same guidance is published for crawlers at /llms.txt, and the MCP server sends it to every connected assistant as its instructions. Method definitions are on the methodology page and in the glossary.

Terms of use

Equity Rank is operated by Leek Ventures LLC, which is not a registered investment adviser, broker-dealer, or financial planner. Everything the endpoints return is algorithmic model output for informational and educational purposes — not investment advice and not a forecast of results. Inputs come from third-party market-data providers and company filings; model logic and scores are Equity Rank's own. Access is rate limited, provided as-is, and subject to the terms of service. For higher limits or questions: [email protected].

About · Methodology · Terms of service · Privacy policy · FAQ