# Agentic Data Mining

> Agentic Data Mining is an institutional data vendor that serves public-record datasets to AI agents over the Model Context Protocol (MCP) as structured JSON. 28 datasets are live today across 7 sectors. Instead of scraping HTML, an agent connects one MCP server and calls typed tools (for example `sec.filings`, `contracts.awards`, `ip.filings`) with shared query parameters and gets clean JSON records back. Coverage figures on each dataset page are computed live from the production ingestion pipeline. Access is licensed by a flat subscription (not free / not metered per query).

Base URL: https://agenticdatamining.ai  ·  MCP endpoint: `https://mcp.agenticdatamining.ai/api/mcp` (header `X-Api-Key: $ADM_API_KEY`)

## How to query

- Add the MCP server to your agent config: `{"mcpServers":{"agentic-data-mining":{"url":"https://mcp.agenticdatamining.ai/api/mcp","headers":{"X-Api-Key":"${ADM_API_KEY}"}}}}`
- Header-less clients (ChatGPT / Claude custom connectors) use a No-Auth capability URL: `https://mcp.agenticdatamining.ai/api/mcp/k/{admc_token}`
- Each dataset is one MCP tool (the `tool` name below). Tools share parameters and return JSON; `limit` defaults to 25 (max 50).
- One-click Claude Desktop bundles (`.mcpb`) are linked from each dataset page.
- Full connection + tool reference: [https://agenticdatamining.ai/docs](https://agenticdatamining.ai/docs)

## Live datasets

### Financial & Government Contracts

- [SEC Form 3/4/5 insider transactions](https://agenticdatamining.ai/datasets/sec-filings) — tool `sec.filings`: Section 16 insider transactions (Forms 3/4/5) from SEC's quarterly bulk extract. Filter by ticker, CIK, insider, transaction code, or trade value.
- [USASpending federal contract awards](https://agenticdatamining.ai/datasets/contracts-awards) — tool `contracts.awards`: Federal contract awards, IDVs, and notices from USASpending.gov (Treasury), joined to awardee entity records. Query by NAICS, agency, awardee, vehicle, or dollar band as awards post.
- [Federal assistance awards](https://agenticdatamining.ai/datasets/gov-assistance) — tool `gov.assistance`: Awarded federal financial assistance from USASpending.gov — grants, loans, direct payments, insurance, and other assistance. Query by recipient, UEI, CFDA program, assistance category, agency, state, or amount. Joins to federal contracts and NIH grants by recipient UEI.
- [DLA DIBBS micro-purchases](https://agenticdatamining.ai/datasets/dla-dibbs) — tool `dla.dibbs`: DLA DIBBS micro-purchases — the sub-$25k Defense Logistics Agency buys USASpending misses. Open solicitations (RFQs) plus award history: last-sold TOTAL award price by NSN, the awardee CAGE, and the date. Query by FSC, NSN, part number, CAGE, or posted date over MCP.
- [DLA FLIS / PUB LOG item catalog](https://agenticdatamining.ai/datasets/dla-flis) — tool `dla.flis`: DLA FLIS / PUB LOG public catalog item intelligence — item identity, FSC, unit of issue, and a catalog reference price for every releasable National Stock Number, keyed by NIIN. Query by niin, nsn, fsc, fsg, or nomenclature over MCP.
- [USPTO patents](https://agenticdatamining.ai/datasets/ip-filings) — tool `ip.filings`: US patent applications and grants by applicant, patent/application number, inventor, CPC class, or filing/grant-date range. Assignee ownership, forward citations, in-force status.
- [SEC company fundamentals (XBRL)](https://agenticdatamining.ai/datasets/sec-fundamentals) — tool `sec.fundamentals`: Standardized XBRL financial fundamentals -- one row per company per fiscal period: revenue, net income, EPS, balance sheet, cash flow. Query by ticker or CIK.
- [Exclusions & debarment](https://agenticdatamining.ai/datasets/risk-exclusions) — tool `risk.exclusions`: SAM.gov government-wide exclusions plus the HHS-OIG LEIE (Medicare/Medicaid debarment), unified in one screening table. Query by name, NPI, UEI, CAGE code, or exclusion program. Flipped live 2026-07-31.
- [DLA prior-award unit prices](https://agenticdatamining.ai/datasets/dla-prior-prices) — tool `dla.prior_prices`: What DLA paid per unit before, for a given NSN -- prior unit price, award date, winning CAGE, quantity and contract number, parsed from the procurement-history block DLA prints on its own solicitation documents. Secondhand by construction (price_basis published_prior_unit), never presented as an observed transaction. Flipped live 2026-07-31; coverage still early relative to a full backfill.
- [Known suppliers by NSN (SAM.gov entity registration)](https://agenticdatamining.ai/datasets/sam-entity-registration) — tool `dla.nsn_suppliers`: Which companies have PREVIOUSLY BEEN AWARDED a given DLA National Stock Number or Federal Supply Class, per DLA's own captured DIBBS award history, each joined by CAGE code to its SAM.gov entity registration -- legal business name, UEI, registration currency, self-reported NAICS codes, entity structure, location, and the SBA small-business size standard for the vendor's primary industry. Answers who has supplied this before, never who could supply it: no NAICS-to-FSC crosswalk exists, so no capability inference is made. Monthly full-refresh snapshot; registration currency is stated as of the loaded snapshot vintage, never against the wall clock. No prices and no debarment screening are returned. Flipped live 2026-08-08.

### Trade & Logistics

- [FMCSA Motor Carrier Census](https://agenticdatamining.ai/datasets/fmcsa-carriers) — tool `trade.carriers`: FMCSA Motor Carrier Census — every registered US motor carrier by USDOT number: legal/DBA name, location, operating authority and status, fleet size, and safety rating. Query by DOT number, name, state, or safety rating.

### Energy & Natural Resources

- [USGS National Water Use](https://agenticdatamining.ai/datasets/usgs-water-use) — tool `env.water_use`: The canonical US water-withdrawal dataset from the US Geological Survey — county and state withdrawals by use category (public supply, irrigation, thermoelectric, industrial, mining, and more) with the ground/surface × fresh/saline breakdown in Mgal/d, on a 5-year compilation cadence (latest complete 2015). Query by state, county, FIPS, category, or year.

### Corporate, Legal & Labor

- [ClinicalTrials.gov registrations](https://agenticdatamining.ai/datasets/health-trials) — tool `health.trials`: Clinical trial registrations by sponsor, condition, status, phase, or intervention. Enrollment, outcomes, recruiting/results status, updated nightly from ClinicalTrials.gov.
- [FDA warning letters](https://agenticdatamining.ai/datasets/fda-warning-letters) — tool `fda.warning_letters`: Search all 964 FDA warning letters -- pharma, telehealth, food, devices, clinical trials. Company, category, resolution status, cited products. Not peptide-specific.
- [NIH research grants](https://agenticdatamining.ai/datasets/nih-grants) — tool `health.nih_grants`: NIH RePORTER research grants -- who funds research on a topic and which orgs get NIH money, plus an SBIR/STTR early-stage biotech signal. Query by org or PI.
- [FDA drug approvals](https://agenticdatamining.ai/datasets/fda-approvals) — tool `fda.approvals`: FDA drug and biologic approvals from openFDA Drugs@FDA -- NDA/ANDA/BLA filings, sponsor, brand/generic name, and an is_novel flag for new molecular entities.
- [SAM.gov Contract Opportunities](https://agenticdatamining.ai/datasets/sam-opportunities) — tool `gov.opportunities`: SAM.gov Contract Opportunities — every federal solicitation, sources-sought, combined synopsis, and award notice over $25,000: agency, NAICS/PSC, set-aside, place of performance, response deadline, and award. Query by NAICS, agency, set-aside, state, or keyword.
- [NPPES healthcare providers](https://agenticdatamining.ai/datasets/healthcare-providers) — tool `health.providers`: Every US healthcare provider in CMS's NPI Registry — identity, practice location, taxonomy, and OIG exclusion status. Query by NPI, name+state, or taxonomy+state over MCP.

### Commerce & Retail

- [Peptide retail pricing](https://agenticdatamining.ai/datasets/peptide-pricing) — tool `commerce.peptides`: Direct-to-consumer peptide retail pricing scraped across hundreds of vendor storefronts (311 tracked, ~200 active) into one normalized schema — compound, vendor, price, currency, package size. Query by compound, vendor, or price.

## Products

- [PeptideIQ](https://agenticdatamining.ai/peptideiq): B2B price data subscription for online research-compound retailers — compare your public prices, catalog and trust signals with competitors.
- [The Peptide Index](https://www.peptidepricehistory.com/): public peptide price-history site built on the peptide retail pricing dataset above.

Operated by Voltairera, LLC d/b/a Agentic Data Mining.

## Site

- [Data catalog](https://agenticdatamining.ai/catalog): all 28 live datasets, filterable by sector and tier.
- [Sectors](https://agenticdatamining.ai/sectors): the 7 sectors and the coverage thesis for each.
- [Coverage](https://agenticdatamining.ai/coverage): live record counts and freshness per dataset.
- [Docs](https://agenticdatamining.ai/docs): MCP quickstart, auth, tool reference, error codes.
- [Pricing](https://agenticdatamining.ai/pricing): flat-subscription tiers.
- [About](https://agenticdatamining.ai/about): who runs it.

## Notes

- Data is normalized US public-record material (SEC/EDGAR, SAM.gov, USPTO, FDA, FMCSA, NIH, NPPES, state registries, and more).
- Further datasets are catalogued as coming-soon and are not yet queryable; only the 28 live tools return data.
- Full per-dataset detail for LLMs: [https://agenticdatamining.ai/llms-full.txt](https://agenticdatamining.ai/llms-full.txt)