Where FPDS sits in the data pipeline
The three systems developers conflate, in order of the contract lifecycle:
- SAM.gov (pre-award): opportunities are posted, vendors are registered. If you want open solicitations, this is the source (see the SAM.gov API guide).
- FPDS (at-award): the contracting officer reports the action, generally within three business days. Award notices you see elsewhere trace back here.
- USASpending (post-award publication): FPDS data, joined with assistance awards and account data, republished with a real API (see the USASpending API guide).
Pulling from the ATOM feed
FPDS’s programmatic interface is an ATOM (XML) feed with fielded queries. Ten records per page, paginated with start:
# contract actions for a vendor UEI, most recent first
curl -s "https://www.fpds.gov/ezsearch/FEEDS/ATOM?FEEDNAME=PUBLIC&q=VENDOR_UEI:%22XXXXXXXXXXXX%22" \
| xmllint --format - | head -60
# actions signed in a date range under a NAICS code
curl -s "https://www.fpds.gov/ezsearch/FEEDS/ATOM?FEEDNAME=PUBLIC&q=PRINCIPAL_NAICS_CODE:%22541512%22+SIGNED_DATE:[2026/01/01,2026/07/01]&start=0"- Useful fields:
VENDOR_UEI,AGENCY_CODE,CONTRACTING_OFFICE_ID,PRINCIPAL_NAICS_CODE,PIID,SIGNED_DATEranges. - The XML uses versioned schemas and namespaces; parse defensively because field availability varies by action type and year.
- For one-off research, skip the feed: ezSearch on fpds.gov exports CSV directly.
What FPDS is good for
- Incumbent watching: new task orders and mods on a competitor’s vehicles appear here first.
- Recompete timing: period-of-performance end dates across active actions are the raw material for forecasting what recompetes in 12 to 18 months.
- Buying-office intelligence: which contracting offices actually sign for the work you sell, not just which agencies.
The honest cost of building on FPDS
Ten records a page of versioned XML, DoD’s 90-day publication delay, schema drift across years, and entity resolution against SAM.gov registrations: a durable FPDS pipeline is real engineering. The CLEATUS API runs that pipeline continuously and exposes the result as market intelligence (incumbents, buying offices, award histories, recompete forecasts) alongside live opportunities, over REST and MCP for AI agents.
