Use the FinFeedAPI SEC MCP server to search filings, extract item text, download filing files, and convert XBRL into JSON.

SEC API MCP

The FinFeedAPI SEC API is available through a hosted Model Context Protocol (MCP) server at https://api.sec.finfeedapi.com/mcp. It exposes the SEC API as self-describing tools that can be used from clients such as Cursor, Claude Desktop, or custom agent runtimes.

Use your regular FinFeedAPI API key in the X-APIKey header.

What is available through MCP?

The SEC API MCP server currently exposes 6 tools grouped into three areas:

Area	What you can do
Filing discovery	Query filing metadata with structured filters or run full-text searches across filing contents.
Filing extraction	Extract the entire classified filing or only a specific item from supported forms such as `8-K`, `10-K`, and `10-Q`.
Raw file and XBRL access	Download filing attachments from EDGAR and convert XBRL filings into JSON.

Typical workflows

Find filings by company or form

Call filings_query when you already know the company identifier such as ticker or cik.
Filter by form_type, filing date, report date, or pagination controls.
Use the returned accession number in extraction and download tools.

Search filing text

Call fulltext_search with text_contains.
Optionally narrow results by form_type and filing date range.
Use the returned filing accession number for follow-up extraction.

Extract filing content

Call extractor_extract_filing to retrieve the full classified filing content.
Call extractor_extract_item when you only need a specific section such as 1.01 from an 8-K.
Use type: "html" when you need markup-preserved output instead of plain text.

Work with source files and XBRL

Call download_file to fetch a specific file from an EDGAR filing package.
Call xbrl_convert to transform XBRL into JSON using either an accession number, an HTML filing URL, or an XBRL XML URL.

Important notes

xbrl_convert accepts three mutually exclusive inputs: htm_url, xbrl_url, or accession_no. Provide only one.
extractor_extract_item expects form-specific item identifiers such as 1.01 for 8-K or 1 for 10-K and 10-Q.
download_file returns file content as a base64-encoded string together with metadata.
Both filings_query and fulltext_search support pagination with page_size and page_number.
The schema currently spells the filing date filters as filling_date_start and filling_date_end. Use those exact parameter names when calling the MCP tools.