SAM.gov tells you what the government wants to buy. It doesn't tell you who won. Once a contract is signed, the award goes into the Federal Procurement Data System (FPDS), the official record of every federal contract action, run by GSA. FPDS has a public ATOM feed that needs no API key and no login.
For a lot of businesses the award is the better lead:
- Subcontractors and suppliers. A prime that just won a $1.4M HVAC job at a VA hospital needs electricians, equipment and materials now.
- Teaming partners. Firms that keep winning in your NAICS code are the ones to team with on the next bid.
- B2B sales. Staffing, bonding, insurance, payroll and software vendors want companies that just landed new government revenue.
- Market research. Who wins in your niche, from which agencies, and for how much.
About 2,300 new awards (base awards, not modifications) are signed on a typical weekday.
The query
The FPDS feed takes a Lucene-style query string. These fields do most of the work:
| Field | Example | Meaning |
|---|---|---|
SIGNED_DATE |
[2026/10/28,2026/11/04] |
Date range the award was signed |
MODIFICATION_NUMBER |
0 |
Only new awards, not modifications |
VENDOR_ADDRESS_STATE_CODE |
TX |
Winner's state |
PRINCIPAL_NAICS_CODE |
23* |
NAICS prefix (wildcards work) |
TYPE_OF_SET_ASIDE |
SDVOSBC |
Set-aside code |
OBLIGATED_AMOUNT |
[100000,] |
Dollar range |
One gotcha: if you repeat a field (VENDOR_ADDRESS_STATE_CODE:TX VENDOR_ADDRESS_STATE_CODE:OK), the values are ANDed, so you get nothing back. Run one query per value instead.
The feed returns 10 entries per page. Page through it with start=0, 10, 20... until a page comes back empty.
The script
Here are new construction awards (NAICS 23) to Texas companies in the last 7 days:
import requests, xml.etree.ElementTree as ET
from datetime import date, timedelta
NS = {"a": "http://www.w3.org/2005/Atom", "ns1": "https://www.fpds.gov/FPDS"}
end = date.today(); start = end - timedelta(days=7)
q = (f"SIGNED_DATE:[{start:%Y/%m/%d},{end:%Y/%m/%d}] MODIFICATION_NUMBER:0 "
"VENDOR_ADDRESS_STATE_CODE:TX PRINCIPAL_NAICS_CODE:23*") # new construction awards to Texas companies
def text(node, path):
el = node.find(path, NS)
return el.text.strip() if el is not None and el.text else None
rows, offset = [], 0
while True:
r = requests.get("https://www.fpds.gov/ezsearch/FEEDS/ATOM",
params={"FEEDNAME": "PUBLIC", "q": q, "start": offset}, timeout=60)
entries = ET.fromstring(r.content).findall("a:entry", NS)
if not entries:
break
for e in entries:
award = e.find(".//ns1:award", NS)
if award is None:
award = e.find(".//ns1:IDV", NS)
if award is None:
continue
rows.append({
"winner": text(award, ".//ns1:vendorName"),
"amount": float(text(award, ".//ns1:obligatedAmount") or 0),
"signed": (text(award, ".//ns1:signedDate") or "")[:10],
"agency": (award.find(".//ns1:contractingOfficeAgencyID", NS).get("name") or "").title(),
"naics": text(award, ".//ns1:principalNAICSCode"),
"city": text(award, ".//ns1:vendorLocation/ns1:city"),
"what": text(award, ".//ns1:descriptionOfContractRequirement"),
})
offset += 10
print(len(rows), "new construction awards to Texas companies in the last 7 days")
for x in sorted(rows, key=lambda x: -x["amount"])[:10]:
print(f'{x["amount"]:>12,.0f} {x["signed"]} {x["winner"]} ({x["city"]}) {x["agency"]} NAICS {x["naics"]}')
Output on Nov 4, 2026:
10 new construction awards to Texas companies in the last 7 days
3,295,000 2026-11-03 GENTRY CONTROLS & CONSULTING LLC (LAGO VISTA) Usda, Departmental Administration NAICS 238210
1,418,449 2026-10-30 DBC-MCA JV, LLC (SAN ANTONIO) Veterans Affairs, Department Of NAICS 238220
416,784 2026-10-28 OMEGA BELT BUILT JOINT VENTURE, LLC (SAN ANTONIO) Smithsonian Institution NAICS 236220
110,150 2026-11-01 HEALING HANDS HEALING HEARTS L.L.C. (AUSTIN) Veterans Affairs, Department Of NAICS 238990
69,324 2026-10-28 PATTON CONTRACTORS INC (FORT WORTH) Public Buildings Service NAICS 236220
60,498 2026-11-02 VALCORP ENTERPRISES LLC (HALTOM CITY) Veterans Affairs, Department Of NAICS 238220
6,750 2026-10-30 SUPERIOR ELEVATOR SERVICES, L.L.C. (PLANO) Federal Prison System / Bureau Of Prisons NAICS 236220
...
Some awards show $0 obligated. These are usually IDIQ or schedule contracts, where the money is obligated later through task orders. Filter on amount > 0 if you only want funded work.
Things to watch
- Sole proprietors. Some awards go to individuals, such as a self-employed contractor or a researcher. If you're building a B2B list, drop records whose vendor name looks like a person's name and has no LLC/Inc/Corp-style marker.
- Contracting officer emails. The raw XML includes the names and emails of the government staff who entered the record. You don't need them for a lead list, so leave them out.
- Reporting lag. Agencies have a few days to report, and DoD awards are published with a 90-day delay. Run with a rolling 7-day window and dedupe on the award ID.
-
Feed changes. GSA has announced that it is moving FPDS reporting into SAM.gov. If the ATOM feed goes away, USAspending.gov's
spending_by_awardAPI (also keyless) carries the same awards, just more slowly.
The no-code version
If you don't want to maintain the parser, paging, name filtering and fallback yourself, I packaged all of it as an Apify Actor: Federal Contract Awards. Each award comes back as one flat record with the winner, UEI, CAGE code, amount, agency, contracting office, NAICS/PSC with descriptions, set-aside, competition type, the winner's business address and phone, small-business/veteran/women-owned/8(a)/HUBZone flags, and a USAspending link. Filter by state, NAICS, set-aside, agency, keyword and amount, then export to CSV or Excel or schedule it daily. It costs $0.90 per 1,000 awards, and Apify's free plan covers about 5,500 a month.
It pairs with the SAM.gov opportunities scraper: opportunities show what's open, and awards show who won it.
A free sample CSV is at loomwise.app/data/government-contract-awards.
Top comments (0)