MCP App Store
by KeenEthics
datastandard icon

datastandard MCP Server for Claude and ChatGPT

By Varshney Projects LLP

Overview

datastandard gives your AI the filed record of India's listed companies on the NSE and BSE: reported financial figures from each company's machine-readable filings; the pages of annual reports, results, investor presentations and offer documents; earnings-call transcripts, page by page; shareholding patterns; corporate actions; exchange announcements; and daily closing prices. Every figure comes with the filing, period and basis it was reported on, so each number in your analysis can be checked against its source. Pages come back with their printed text, and as images for an AI that reads them, so a figure can be read off the printed page. You can also save the reports, models and memos your AI writes to your firm's private store and share them with colleagues. datastandard does not give investment advice, recommendations, ratings or live prices. You need a datastandard account, which you can create when you connect.

How to connect

  1. Claude: Open the datastandard listing in the Claude connectors directory and choose Connect. Claude listing datastandard exposes 39 tools here.
  2. ChatGPT: Open the datastandard listing in the ChatGPT app directory and choose Connect. ChatGPT listing datastandard exposes 39 tools here.

Tools

confirm_artefact

ChatGPT
The second half of a save, after the file has been uploaded to the link save_artefact gave. Until then nothing is saved. Safe to call twice.

delete_artefact

ChatGPT
Removes a saved artefact for everyone in the firm, including the person who saved it. It cannot be restored from here.

get_action_pages

ChatGPT
The company's own filing behind one corporate action, as page images: the board outcome, the public announcement, the record-date notice. The action_id comes from get_corporate_actions.

get_announcement_facts

ChatGPT
Every field of the tagged form filed with an announcement — the date, venue and attendees of an analyst meeting, the dividend rate a board recommended — as the company tagged them, in filing order. Only rows whose form is xbrl carry one; the announcement_id comes from list_announcements.

get_announcement_pages

ChatGPT
The company's own filing behind one announcement, as page images, in page order — the letter, the press release, the results, the deck or the transcript the exchange listed that day. The announcement_id comes from list_announcements.

get_artefact

ChatGPT
One saved artefact with a short-lived signed link to the file, readable by fetching the link.

get_call_pages

ChatGPT
Consecutive pages of one earnings-call transcript by page number, as images. Does not search: search_call_text finds the page and this reads the exchange around it. One company, one fiscal year, one quarter per call.

get_company_profile

ChatGPT
Identity, listing status, market cap and industry classification for one company.

get_corporate_actions

ChatGPT
Every dividend, bonus, split, buyback, rights issue, demerger and meeting from the exchanges' record-date register, with ex-date and record date. Where terms_in_filing is true, get_action_pages reads the price, size or ratio off the company's own filing.

get_document_download_url

ChatGPT
A short-lived link to one filing. Metered far more heavily than a fact lookup or a page.

get_document_meta

ChatGPT
Details of one filing: its type, the period it covers, and its file.

get_filing_pages

ChatGPT
Specific pages of an annual report by PDF page number, each as an image of the printed page. Does not search: search_filing_text finds the page and this reads around it. One fiscal year per call. Each page carries its printed page number, the citation.

get_filing_section

ChatGPT
The balance sheet, profit and loss, cash flow or changes in equity from an annual report, as page images, by section name. One fiscal year per call. Where a statement is not mapped, search_filing_text finds it.

get_formula

ChatGPT
Formulas for common financial ratios (growth, margins, returns, leverage, cash flow, valuation, banks, NBFCs, insurers, ownership, prices), each with the filed figures it uses, the companies it applies to and when it does not apply. Counts towards the per-minute limit, never the plan's call allowance.

get_offer_document_pages

ChatGPT
Specific pages of one offer document by page number, as images. Does not search; the document_id comes from search_offer_document_text or list_documents.

get_pages

ChatGPT
The page images and text of one filing, by document_id and page numbers, for any document search_pages returns. A figure on a filing page is read from the image.

get_presentation_pages

ChatGPT
Consecutive slides of one investor presentation by page number, as images. Does not search. With no quarter given, the deck that reports the full year is read.

get_segments

ChatGPT
Segment revenue, result and assets as the company reported them, on one basis and one period length per answer. Rows flagged reconciliation or residual are not part of a segment total.

get_shareholding

ChatGPT
Who owns the company, from its SEBI Regulation 31 filings: promoter and public holding, institutions, named large holders, promoter pledge. Snapshots are as at the date on the row. Percentages are already percentages. The category rows form a tree; the parent is the total.

get_stock_prices

ChatGPT
Open, high, low, close and volume for each trading day, newest first, as published and not adjusted for bonuses or splits (scale_changes lists any inside the window). Not a live quote. With no dates given, the most recent year. Counts towards the per-minute limit, never the plan's call allowance.

get_xbrl_facts

ChatGPT
Filed financial figures for one company, one row per line item with its value, period, basis and filing. Consolidated by default. The element filter ("revenue", "profit after tax", "EPS") narrows an otherwise large answer. Lines with similar names are different figures.

guide

ChatGPT
What datastandard is, what it is not, what its data carries, and the guided recipes. Counts towards the per-minute limit, never the plan's call allowance.

list_announcements

ChatGPT
Every filing the company made with its exchange — board outcomes, analyst meetings, acquisitions, record dates, credit ratings — newest first, with the exchange's own subject and category and the law's name for the event. get_announcement_pages reads the filing behind a row; get_announcement_facts reads a tagged form.

list_artefacts

ChatGPT
The reports, models and memos this firm has saved, filtered by company, tag or a search over titles, descriptions and tags. The search does not reach inside the files.

list_companies

ChatGPT
The covered companies, optionally filtered by sector or minimum market cap (in crore). Paged; pagination.truncated says whether more follow.

list_documents

ChatGPT
The filings available for a company: annual reports, results, shareholding filings, earnings calls, investor presentations, announcements, corporate-action filings and offer documents. Paged. Returns the document_id the other tools take.

list_screen_fields

ChatGPT
The exact names screen_companies accepts: line items as get_xbrl_facts prints them, shareholding categories and measures, the exchange's own filing categories with their event_ref and the day each was last filed, sectors and industries, segment measures, and filed declarations with every value filed.

ping

ChatGPT
Whether datastandard is reachable and the connection is valid, proven by a live query for company data. Counts towards the per-minute limit, never the plan's call allowance.

save_artefact

ChatGPT
Saves a produced file (report, model, memo, deck) into the firm's store. Saving is two calls with an upload in between: this returns an upload link for an HTTP PUT of the file, and confirm_artefact completes it. Nothing is saved until confirmed. The file is stored, never read.

screen_companies

ChatGPT
The covered companies whose filed figures, filed declarations, shareholding, segments and exchange filings meet a set of conditions, with the filed values that decided each result and the filing behind every value. Up to 50 companies a page; companies without a filed value to test are listed apart.

search_announcement_text

ChatGPT
Where something appears across a company's announcement filings, as pages, best match first. Each page names the announcement it belongs to and the date it was filed.

search_call_text

ChatGPT
Where a subject was discussed on a company's earnings calls, as transcript pages: one fiscal year (quarter optional), a range of dates, or the latest few calls. The speaker's name is printed on the page beside their words; datastandard does not label speakers. A page matches when it carries most of the query words.

search_companies

ChatGPT
The companies matching a name, NSE symbol, ISIN or BSE code. When several match, the candidates are returned and none is picked.

search_filing_text

ChatGPT
Where something is discussed in one annual report for one fiscal year, as pages. A page matches when it carries most of the query words; a quoted phrase must appear exactly. Each page says which basis it belongs to.

search_offer_document_text

ChatGPT
Where something appears in a company's offer documents, as pages. Each page says which of the three it came from: the draft carries no price, the red herring usually prints a placeholder for the band, the prospectus carries the struck price.

search_pages

ChatGPT
Filing pages that discuss a subject across up to 50 companies and a range of fiscal years, from annual reports, earnings calls, investor presentations, exchange announcements, offer documents and corporate-action filings. Each hit carries the page text and where it is printed.

search_presentation_text

ChatGPT
Where something appears on a company's investor presentation slides for one fiscal year, as slides. The extracted slide text has no layout, so a figure is read off the slide image (images true, or get_presentation_pages). Each page says which kind of deck it came from.

share_artefact

ChatGPT
Gives colleagues access to a saved artefact: the whole firm, or named people by login email. They can read and download it, not change or delete it.

unshare_artefact

ChatGPT
Withdraws access to a saved artefact that was shared: from the whole firm, or from named colleagues by login email. The artefact itself is not changed or deleted.

confirm_artefact

Claude

delete_artefact

Claude

get_action_pages

Claude

get_announcement_facts

Claude

get_announcement_pages

Claude

get_artefact

Claude

get_call_pages

Claude

get_company_profile

Claude

get_corporate_actions

Claude

get_document_download_url

Claude

get_document_meta

Claude

get_filing_pages

Claude

get_filing_section

Claude

get_formula

Claude

get_offer_document_pages

Claude

get_pages

Claude

get_presentation_pages

Claude

get_segments

Claude

get_shareholding

Claude

get_stock_prices

Claude

get_xbrl_facts

Claude

guide

Claude

list_announcements

Claude

list_artefacts

Claude

list_companies

Claude

list_documents

Claude

list_screen_fields

Claude

ping

Claude

save_artefact

Claude

screen_companies

Claude

search_announcement_text

Claude

search_call_text

Claude

search_companies

Claude

search_filing_text

Claude

search_offer_document_text

Claude

search_pages

Claude

search_presentation_text

Claude

share_artefact

Claude

unshare_artefact

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/datastandard"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/datastandard)