Primary Sources

io.github.simonmak-ascentv1.1.1Updated Oct 3, 2026

Free primary-source APIs across 8 categories, with provenance envelopes.

VerifiedStreamable HTTPWeb executableWeb Search & ScrapingData & AnalyticsKnowledge & Memory

Overview

AI-generated overview

Gives an assistant 50 read-only tools over free public primary-source APIs, from SEC filings and company registries to macro, news, research and open data.

What it does
Exposes 50 tools across eight categories: filings (SEC EDGAR, HKEXnews), registries (UK Companies House, GLEIF, ACRA, EU VIES), macro (World Bank, FRED, HKMA, DBnomics, ECB, US Treasury, UN, WHO), news (GDELT, Hacker News, Wikimedia pageviews), open data (DATA.GOV.HK, CKAN, data.europa.eu), research (Crossref, OpenAlex, arXiv, PubMed, Europe PMC, DataCite, DOAJ, Zenodo, Semantic Scholar), knowledge (Wikidata, Wikipedia, Open Library, Internet Archive, geocoding, weather, earthquakes, GBIF, PubChem, openFDA, ClinicalTrials.gov, NASA APOD) and tools (MCP Registry search, Jina Reader). Each result is wrapped in a provenance envelope with source, retrievedAt and data so the assistant can cite…
When to use it
Useful when an assistant needs citable primary sources rather than general web answers: company filings and registry records, economic and development indicators, scholarly literature, official open-data catalogues, or reference facts. Most tools work without any key, so it is easy to try.
Requirements
Runs as a remote Streamable HTTP endpoint or locally via npx with Node.js 20 or newer. No runtime dependencies. 47 of 50 tools need no key; COMPANIES_HOUSE_API_KEY enables two UK Companies House tools and FRED_API_KEY enables one FRED tool. NCBI_API_KEY and NASA_API_KEY are optional rate-limit boosts, and RESEARCH_CONTACT is an optional politeness contact string. Network access to the upstream public APIs is required.
Before you install
All tools appear read-only and no authentication is declared for the server itself. Optional secrets COMPANIES_HOUSE_API_KEY, FRED_API_KEY, NCBI_API_KEY and NASA_API_KEY are sent to the server or process, so treat them as credentials. Queries and fetched URLs go to third-party public APIs, and source licences vary: Wikipedia is CC BY-SA and needs attribution, while HKEXnews, GDELT and SEC filings are fetched on demand and not redistributed.

Installation

In SourceWeft

  1. Open Primary Sources in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "primary-sources-mcp": {
      "type": "http",
      "url": "https://primary-sources-mcp.vercel.app/mcp"
    }
  }
}

README

primary-sources-mcp

Free primary-source APIs for AI agents. 50 tools across 8 categories — filings, registries, macro, news, open-data, research, knowledge and tools. SEC EDGAR, HKEXnews, World Bank, GDELT, DATA.GOV.HK, HKMA, the official MCP Registry, UK Companies House, FRED and Jina Reader — as MCP tools, each result wrapped in a { source, retrievedAt, data } provenance envelope.

[CI] [Secret Scan] [MCP Tool Definition Quality] [MCP Registry] [Glama] [License: MIT]

Overview

Zero runtime dependencies. Node >= 20. 50 tools, 47 of them keyless.

Quick start

json
{  "mcp": {    "primary-sources": {      "type": "local",      "command": ["npx", "-y", "@simonmak-ascent/primary-sources-mcp"],      "enabled": true,      "timeout": 120000    }  }}

Hosted (Streamable HTTP): https://primary-sources-mcp.vercel.app/mcp.

Tools by category

filings

ToolSourceKey
sec_edgar_fulltextSEC EDGAR full-text search (US filings)none
sec_edgar_companySEC EDGAR registrant + recent filingsnone
hkex_announcementsHKEXnews listed-company announcement search (HK)none
hkex_companyHKEXnews listed-company resolve + recent filingsnone

registries

ToolSourceKey
companies_house_searchUK Companies House searchCOMPANIES_HOUSE_API_KEY
companies_house_companyUK Companies House recordCOMPANIES_HOUSE_API_KEY
gleif_leiGLEIF Legal Entity Identifier records (CC0)none
acra_sg_companyACRA Singapore corporate registry (data.gov.sg)none
vies_vatEU VIES VAT validationnone

macro

ToolSourceKey
world_bank_indicatorWorld Bank development indicatorsnone
fred_seriesSt. Louis Fed FRED seriesFRED_API_KEY
hkmaHong Kong Monetary Authority public APInone
dbnomics_seriesDBnomics (World Bank, IMF, OECD, Eurostat, ECB, BIS, …)none
ecb_seriesECB Data Portal (SDMX-JSON)none
treasury_fiscalUS Treasury Fiscal Datanone
un_sdgUN Sustainable Development Goal indicatorsnone
un_comtradeUN Comtrade trade statisticsnone
who_ghoWHO Global Health Observatorynone

news

ToolSourceKey
gdelt_newsGDELT DOC 2.0 global newsnone
hn_searchHacker News (Algolia)none
wikimedia_pageviewsWikimedia pageview metrics (CC0)none

open-data

ToolSourceKey
hk_open_data_searchDATA.GOV.HK catalogue searchnone
hk_open_data_filterDATA.GOV.HK filter APInone
ckan_searchCKAN portals (data.gov.uk, data.gov.au)none
data_europa_searchdata.europa.eu EU open-data cataloguenone

research

ToolSourceKey
crossref_worksCrossref DOI metadatanone
openalex_worksOpenAlex scholarly worksnone
arxiv_searcharXiv preprintsnone
pubmed_searchPubMed (NCBI E-utilities)none (optional NCBI_API_KEY)
europepmc_searchEurope PMCnone
datacite_doisDataCite research-data DOIsnone
doaj_articlesDirectory of Open Access Journalsnone
zenodo_recordsZenodo research outputsnone
semantic_scholar_worksSemantic Scholar (rate-limited keyless)none

knowledge

ToolSourceKey
wikidata_searchWikidata entities (CC0)none
wikipedia_summaryWikipedia summaries (CC BY-SA)none
open_library_searchOpen Library booksnone
gutendex_booksProject Gutenberg via Gutendexnone
internet_archive_searchInternet Archivenone
world_bank_countryWorld Bank country reference metadatanone
nominatim_geocodeOpenStreetMap Nominatim geocodingnone
open_meteo_forecastOpen-Meteo current weathernone
usgs_earthquakesUSGS earthquake cataloguenone
gbif_speciesGBIF taxonomic backbonenone
pubchem_compoundPubChem compound datanone
openfda_drugopenFDA drug adverse eventsnone
clinicaltrials_studyClinicalTrials.gov studiesnone
nasa_apodNASA Astronomy Picture of the Daynone (optional NASA_API_KEY)

tools

ToolSourceKey
mcp_registry_searchOfficial MCP Registrynone
jina_readJina Reader (URL → markdown)none (free tier)

Environment

VariableRequiredPurpose
RESEARCH_CONTACTnoUser-Agent contact string (politeness for SEC/GDELT/OpenAlex).
COMPANIES_HOUSE_API_KEYfor 2 toolsFree key from Companies House.
FRED_API_KEYfor 1 toolFree key from the St. Louis Fed.
NCBI_API_KEYnoRaises PubMed (NCBI) rate limits.
NASA_API_KEYnoRaises NASA APOD limits (falls back to shared DEMO_KEY).

Provenance & attribution

Every data tool returns text of the form:

json
{ "source": "World Bank Indicators", "retrievedAt": "2026-10-02T16:00:00.000Z", "data": { } }

so an agent can cite the origin and time of every fact.

Source licences vary: GLEIF and Wikidata are CC0; World Bank, Crossref, OpenAlex, DataCite, DOAJ, Zenodo, GBIF, openFDA, ClinicalTrials.gov and US government data are public-domain/open; ACRA is under the Singapore Open Data Licence; Wikipedia is CC BY-SA (attribution required); HKEXnews, GDELT and SEC filings are retrieved link-first and fetched on demand, not redistributed.

Development

bash
npm installnpm run typechecknpm testnpm run buildnode dist/index.js   # stdio MCP server

License

MIT — see LICENSE.

Source: README.md at commit c9c3a95

Tools

0
Tool metadata has not been indexed yet.

Version history

1
  1. v1.1.1LatestOct 3, 2026