Sales & risk Parses
Parses are record-change detection.
Every ParseData feed is a cursor-paginated API over official source files, and every Parse below is the same simple machine: poll a feed with since= on your schedule, dedupe on the stable key, and act when a new record appears or a status field changes. 30 recipes, organized by the team that runs them — each grounded in a specific feed and filter, with one fully worked build per team — plus a healthcare vertical served by our sister estate Healthparse.
Honest mechanics: pull or push, your call. Poll with since= on your own schedule (cron, Clay, Zapier, warehouse job) — or save a Parse in the builder and the gateway checks on your schedule and POSTs new rows to your webhook, HMAC-signed. Either way it’s $0.05–$0.10 per page-call, so a daily Parse costs the same as the daily check it replaces. (Healthcare feeds push via the Healthparse enterprise setup, not through govparse deliveries.)
The pattern
Anatomy of a Parse
01 · First pull
GET the feed with your filters; pass page.next_cursor back as cursor= until page.has_more is false. Up to 1,000 flat rows per page-call.
02 · Increment
On your schedule, re-call with since= (or added_since=) set to your last watermark and drain again. Only the records that changed come back.
03 · Dedupe
Every feed has a stable key — event_key, usdot, subject_key + last_marker_date. Keep the keys you've seen; a republished row never fires twice.
04 · Deliver
POST new rows anywhere HTTP reaches: Slack webhook, CRM API, Google Sheet, a Clay HTTP API column, or a Zapier Webhooks step.
For enterprise sales & gtm
Enterprise sales & GTM
Dated official records that tell you which account to call this week — a new carrier registered in your territory, a competitor's filings naming your technology, a citation with an abatement deadline. Each recipe below is a feed plus a filter; the record itself is the talking point.
Worked example
Build this Parse: H-1B hiring-intent accounts
For sales-ops at infrastructure and enterprise-software vendors: an employer whose new filings name your technology has engaged counsel and filed with DOL — a stronger commit than a job posting.
01First pull
One page-call returns up to 1,000 rows at $0.05. Drain the window by passingpage.next_cursorback ascursor=untilpage.has_moreis false.curl "https://api.govparse.io/v1/feeds/workforce-filings?since=2026-01-01&soc=15-1252&state=WA,CA&limit=1000" \ -H "x-api-key: gp_live_..."02Filter for intent
Keep rows whose free-textjob_titlenames your stack (“Snowflake”, “Kubernetes”, “SAP S/4HANA”) — one keyword filter in your pipeline, or a Clay formula column. Per-account variant:employer={{Company Name}}(punctuation-insensitive).03Poll
Re-run withsince=advanced to your last watermark (the maxevent_dateseen). New rows arrive with each quarterly OFLC disclosure drop, so a weekly poll usually returns empty — and costs a nickel. Dedupe onevent_key(lca:<case number> / perm:<case number>) so no case ever fires twice.04Deliver
Anywhere HTTP reaches. In Clay, add an “HTTP API” column (save the key once as a headers account:x-api-key) and map response fields to columns. In Zapier, pair a Schedule trigger with a “Webhooks by Zapier” GET step. Or run a cron that POSTs new rows to a Slack incoming webhook, your CRM’s REST API, or a Google Sheet. No proprietary connector required — it’s a GET request.
01Machine-guarding equipment vendor — regional sales rep
Enforcement Events feed →- Asks
- “Which manufacturers in my territory were cited under a machine-guarding standard in the last six months, and when is each abatement deadline?”
- The feed returns
- OSHA rows filtered by NAICS and state, each carrying the violated standard codes (e.g. 19100212), current penalty, site address, and abatement_latest_date — the date by which the cited hazard must be corrected.
- Action
- Call into a dated compliance window with a concrete talking point — the citation on record and its deadline — instead of cold-calling.
02Wastewater / effluent-treatment engineering firm — BD lead
Enforcement Events feed →- Asks
- “Which food-processing and metal-finishing plants in my state currently carry a Clean Water Act significant-noncompliance flag?”
- The feed returns
- EPA rows filtered on NAICS prefixes 3116 and 3328 where codes include CWA_SNC, with compliance status, the facility's formal-action count and most recent penalty, and a source_url to the EPA detailed facility report.
- Action
- Pitch the treatment upgrade while the flag is on record and before a consent decree defines the project for you — the highest-margin, least-competitive moment.
03Environmental remediation contractor — territory planner
Enforcement Events feed →- Asks
- “Map every facility with a current RCRA significant-noncompliance flag inside these five industrial ZIP codes.”
- The feed returns
- EPA rows filtered by state and ZIP where codes include RCRA_SNC, each with the full site address and a source_url to the official facility report for verification.
- Action
- Turn scattered compliance flags into a drive-time-optimized prospecting route where every stop has a documented compliance obligation.
04Ventilation / respiratory-protection systems vendor
Enforcement Events feed →- Asks
- “Which sites appear in both an OSHA air-contaminant citation and an EPA Clean Air Act high-priority-violation flag?”
- The feed returns
- OSHA rows citing standards 19101000/19100134 joined via entity_id and address to EPA rows carrying CAA_HPV — the same site on record with two agencies.
- Action
- Sell one capture-ventilation project that addresses both records at once — an ROI pitch grounded entirely in the official record.
05EHS software account executive
Enforcement Events feed →- Asks
- “Which mid-size employers just recorded a repeat inspection on a standard they were previously cited under?”
- The feed returns
- Per-entity violation history via entity_id joins: the same standard code appearing on inspections in successive years, with dates and penalties as recorded.
- Action
- Reach out at the moment the record shows manual corrective-action tracking did not prevent a recurrence — the natural EHS-software conversation.
06Enterprise SaaS / systems-integrator sales team
Workforce Filings feed →- Asks
- “Which companies started filing for 'Snowflake', 'Databricks', or 'SAP S/4HANA' titles in the last two quarters and had zero such titles before?”
- The feed returns
- Rows whose free-text job_title matches your stack keywords, diffed against the same employer's prior-period filings — e.g. six 'Senior Data Engineer - Snowflake/dbt' cases filed at one worksite where none existed in the trailing eight quarters.
- Action
- Route as a high-intent lead to the account exec for the complementary product or migration services — before the RFP goes out.
07Commercial real estate broker / site-selection consultant
Workforce Filings feed →- Asks
- “Show me employers whose latest worksite city is a metro where they had no prior filings — new-location tells before a lease is announced.”
- The feed returns
- worksite_city/state per case, compared against the employer's full historical worksite set: e.g. a California robotics firm's first-ever Ohio filings, eleven cases for controls and automation engineers at a Columbus worksite.
- Action
- Pitch the tenant-rep mandate and comps before the company announces the site or issues an RFP.
08PE / growth-equity deal sourcer
Workforce Filings feed →- Asks
- “Which private companies filed PERM cases for senior, specialized staff for the first time in the last year?”
- The feed returns
- PERM rows filtered to first-time filers, with seniority parsed from job_title and offered wages relative to the SOC's distribution — e.g. four 'Principal ML Engineer' and 'Director, Platform' cases at top-decile wages from a private fintech.
- Action
- Prioritize proprietary outreach before the company runs a formal raise or hits banker radar — PERM filings signal long-horizon commitments.
09Executive / technical recruiter
Workforce Filings feed →- Asks
- “Where are competitors bidding up wages for a specific role right now?”
- The feed returns
- wage_from/wage_to by SOC code, employer, and worksite metro, trended by decision_date — e.g. two rival chip-design firms in Austin whose offered wages for RTL design engineers climbed 25-30% across successive quarters of filings.
- Action
- Pitch a retained search to the employer whose filings show expansion, and source candidates from the one whose filings show it lagging.
10Industrial equipment / capital-goods sales team
Workforce Filings feed →- Asks
- “Which manufacturers are filing for plant, controls, and process engineers at brand-new US worksites?”
- The feed returns
- Cases in SOC 17-2112/17-2141 at worksites new to that employer — e.g. nine manufacturing-process and controls-engineer filings at a first-time Tennessee worksite from an auto-parts maker.
- Action
- Target the site with capital-equipment and automation proposals ahead of the buildout's procurement cycle.
11Used-truck dealer / heavy-equipment auctioneer
Carrier Analytics feed →- Asks
- “Find sizable fleets whose census record went inactive recently, so I can make a cash-for-fleet offer.”
- The feed returns
- Rows where operating_status changed from active, sorted by power_units, with legal name, DBA, and address as filed — a direct-mail-ready list of fleets whose records show a wind-down.
- Action
- Offer liquidity to owners ahead of the auction cycle, buying rolling stock below auction value.
12Staffing / outplacement firm — talent acquisition lead
Distress Signals feed →- Asks
- “Give me a named, dated list of employers with WARN notices effective in the next 60 days near my branches.”
- The feed returns
- marker=warn_notice_180d rows filtered by state, each markers entry carrying worksite city, workers affected, and notice/effective dates — the legally mandated 60-day lead is the recruiting window.
- Action
- Pre-recruit the displaced pool weeks before competitors and pitch outplacement services to the filing employer while the notice is fresh.
13Special-situations PE / distressed-M&A deal sourcer
Distress Signals feed →- Asks
- “Which companies added going-concern language and also have a WARN notice or authority loss on record?”
- The feed returns
- min_markers=2 rows where marker_keys pairs going_concern_language_1y with warn_notice_180d or fmcsa_authority_revocation_180d — auditor-certified doubt corroborated by an operational record, with CIK/EIN for your own diligence joins.
- Action
- Open a proactive restructuring, DIP-financing, or take-private conversation weeks before bankers run a formal process.
14Used-truck dealer / heavy-equipment auctioneer
Distress Signals feed →- Asks
- “Which carriers just had operating authority revoked, so I can offer to buy the now-idle fleet before the repo men arrive?”
- The feed returns
- marker=fmcsa_authority_revocation_180d rows with the USDOT number and docket in the markers detail — join to the carrier-analytics feed for fleet size and address, and check for a co-occurring WARN closure notice confirming the whole operation is winding down.
- Action
- Make a direct cash-for-fleet offer to an owner who needs liquidity, buying rolling stock below auction value.
For credit & risk
Credit & risk
Monitor a book of accounts against the official record on your schedule: authority revocations, insurance filings going to zero, WARN notices, going-concern language. Every alert cites the underlying record with dates, so the escalation memo writes itself.
Worked example
Build this Parse: WARN / distress suppression-and-trigger on your account list
For credit, AR, and portfolio teams — and sales-ops running the suppression side: route accounts carrying an official-record distress marker out of expansion sequences, or into review.
01First pull
Pull the full composite once — about $2.30 for all 23,207 subjects at $0.10 per 1,000-row page. Drain withpage.next_cursoruntilpage.has_moreis false. Narrow withmarker=warn_notice_180d,min_markers=2, orstate=.curl "https://api.govparse.io/v1/feeds/distress-signals?limit=1000" \ -H "x-api-key: gp_live_..."02Match your list
Joindisplay_name+state(andein/cikwhere present) against your accounts. One pull, two plays: suppress matched accounts out of outbound sequences, or trigger a limit / covenant review with the cited records attached.03Poll
Re-run withsince=set to your last run date — the feed’s axis islast_marker_date, so you get new subjects plus subjects whose marker set moved. Dedupe onsubject_key + last_marker_date.curl "https://api.govparse.io/v1/feeds/distress-signals?since=2026-07-16&limit=1000" \ -H "x-api-key: gp_live_..."04Deliver
Anywhere HTTP reaches. In Clay, add an “HTTP API” column (save the key once as a headers account:x-api-key) and map response fields to columns. In Zapier, pair a Schedule trigger with a “Webhooks by Zapier” GET step. Or run a cron that POSTs new rows to a Slack incoming webhook, your CRM’s REST API, or a Google Sheet. No proprietary connector required — it’s a GET request.
Strictly observational: which records exist, with dates — never a score or a 'distressed' label. Not a consumer report; no FCRA-regulated use.
01Special-situations / distressed-asset analyst
Enforcement Events feed →- Asks
- “Which private operators have enforcement records across all three agencies in the last 24 months?”
- The feed returns
- Rows joined on entity_id across agencies — e.g. one resolved entity with repeat OSHA citations, an RCRA_SNC flag on its facility, and WHD back wages recorded, with match tier and confidence disclosed on every link.
- Action
- Build a watchlist from cross-agency records months before anything comparable appears in credit or financial data.
02Workers'-comp underwriter / commercial insurance broker
Enforcement Events feed →- Asks
- “Which insureds on my renewal book have new OSHA inspection activity this year versus their trailing baseline?”
- The feed returns
- A weekly delta of inspection rows matched to your insured schedule by name and address, with inspection type codes, violation counts, and penalties as recorded.
- Action
- Reprice or attach risk-engineering requirements at renewal, while the loss trend is still developing.
03Long/short equity alt-data analyst
Workforce Filings feed →- Asks
- “Which brick-and-mortar retailers show a step-change in robotics and machine-learning titles at distribution-center worksites this year?”
- The feed returns
- Cases filtered on SOC families 17-2xxx / 15-2051 and title keywords, grouped by employer and worksite — a robotics-title count moving from 3 to 41 year over year, concentrated at two fulfillment-center locations.
- Action
- Position ahead of a capex or margin inflection the filings surface quarters before guidance, and size against consensus assumptions.
04Short-seller / forensic research shop
Workforce Filings feed →- Asks
- “Which companies marketing themselves as 'AI-first' have filings that contain no ML or research titles at all?”
- The feed returns
- The employer's complete filed title mix: e.g. a small-cap issuer whose filings on record show zero cases in ML/robotics SOC families while 90% are entry-level analyst titles — as filed with DOL, with case numbers.
- Action
- Document the gap between public narrative and the hiring record in a diligence memo, citing the filings directly.
05Freight factoring underwriter
Carrier Analytics feed →- Asks
- “Re-check every carrier in my factoring book overnight: whose operating status, authority statuses, or insurance filings changed since yesterday?”
- The feed returns
- Your roster joined by USDOT number to the daily census: operating_status flips, active_authorities counts moving, and insurance_active_filings going to zero — the recorded events that precede a carrier going dark.
- Action
- Freeze new advances and accelerate collections on accounts whose official records changed, before open invoices become write-offs.
06Trade-credit insurer / AR team
Carrier Analytics feed →- Asks
- “Across every trucking counterparty we cover, which ones show recorded authority or insurance-filing changes this quarter?”
- The feed returns
- A portfolio-wide diff over the daily feed: per-carrier authority_statuses, insurance filing counts, prior_revoke_flag, and operating status, keyed by USDOT for a clean join to your book.
- Action
- Reprice or adjust limits on the accounts whose official records moved, before a claim materializes.
07Equipment lessor / asset-recovery firm
Carrier Analytics feed →- Asks
- “Which carriers with 20+ power units now show inactive operating status or no active authorities on file?”
- The feed returns
- Census rows filtered by min_power_units where operating_status is inactive or active_authorities is zero, with fleet composition (trucks, buses, power units) and the physical address on file.
- Action
- Prioritize repossession and restructuring outreach where the records indicate equipment is likely idle — while assets are still on the lot.
08Regional bank commercial-loan workout officer
Carrier Analytics feed →- Asks
- “Scan my C&I book nightly and escalate any trucking borrower whose official record changed.”
- The feed returns
- Borrowers matched by USDOT: authority status transitions, insurance filing counts, safety rating changes, and 24-month out-of-service and crash summaries — a nightly monitoring cron, not a one-time pull.
- Action
- Trigger covenant review on documented record changes and tighten credit before a default, improving recovery.
09Freight-tech / insurtech data team
Carrier Analytics feed →- Asks
- “Give me the entire carrier universe as a base table for pricing and risk models — and a scheduled check for new entrants.”
- The feed returns
- All 4.47M census rows in USDOT order for the full-universe pull, then added_since polling for newly registered carriers (the natural shape for new-entrant onboarding funnels), with SaferGrade scores and SMS summaries pre-joined.
- Action
- Skip the five-source ingestion project: land one flat table in your warehouse and refresh deltas daily.
10Freight factoring underwriter / transportation lender
Distress Signals feed →- Asks
- “Which carriers in my factoring book have an official-record distress marker on file right now — and which have more than one?”
- The feed returns
- Rows filtered on marker=fmcsa_authority_revocation_180d matched to your client roster, each with the order type and effective date in the markers detail — plus any WARN notice or enforcement record the same subject carries, with marker_count for triage.
- Action
- Freeze new advances and accelerate collection on open invoices while the carrier is still hauling — turning a likely writeoff into a recovered receivable.
11Trade-credit insurer / supplier AR team
Distress Signals feed →- Asks
- “Which insured buyers were touched by two independent official-record markers inside the same 60-day window?”
- The feed returns
- min_markers=2 rows matched to your insured list: e.g. an FMCSA authority action and a WARN notice recorded weeks apart at the same subject, with per-marker dates so the 'two independent sources' escalation rule is computable from the record itself.
- Action
- Reduce the coverage limit or reprice the policy before a claim materializes, with both source records citable in the file.
12Industrial landlord / equipment lessor — asset manager
Distress Signals feed →- Asks
- “Which of my tenants and lessees carry a plant-closing notice or a revoked license on record?”
- The feed returns
- Your rent roll matched by name and state: subjects whose markers include a WARN closure at the property's city or an authority revocation, with dates — the record, not a prediction, months before a payment default.
- Action
- Start re-marketing the space, secure a letter of credit, or restructure the lease while the counterparty is still current.
13Regional bank commercial-loan workout officer
Distress Signals feed →- Asks
- “Scan my C&I book weekly and escalate any borrower an official record just touched with a distress marker.”
- The feed returns
- A since=last-run pull matched to borrower names: new subjects and subjects whose last_marker_date moved, each with the specific records (WARN, 8-K item, revocation) that changed — an auditable Parse for covenant review.
- Action
- Invoke material-adverse-change review and tighten the credit before default, with the official record attached to the memo.
For compliance & ops
Compliance & ops
Recurring checks over the same feeds: screen suppliers against enforcement records, vet carriers before first tender, watch competitor capability builds. Match confidence is disclosed on every row — you set the threshold.
Worked example
Build this Parse: new carrier registered in your territory
For carrier-onboarding and vetting ops at brokers, insurers, and factors — each new registration arrives with enough of the official record to run your onboarding checklist before first tender. (The same Parse is the classic new-entrant sales play: new carriers need insurance, factoring, fuel cards, and ELDs in their first 90 days.)
01First pull
The census carries anadd_dateper USDOT number;added_since=returns carriers added to the census in your window. $0.05 per 1,000-row page; drain with the cursor as above.curl "https://api.govparse.io/v1/feeds/carrier-analytics?added_since=2026-06-23&state=TX&min_power_units=1&limit=1000" \ -H "x-api-key: gp_live_..."02Poll daily
The census refreshes daily. Advanceadded_since=to your last run date and dedupe onusdot— the stable official identifier — so a carrier republished in a new census vintage never fires twice.03Route on the record
Every row carriesoperating_status,active_authorities,insurance_active_filings, and fleet size — apply your own onboarding criteria to the record as published, and queue what passes.04Deliver
Anywhere HTTP reaches. In Clay, add an “HTTP API” column (save the key once as a headers account:x-api-key) and map response fields to columns. In Zapier, pair a Schedule trigger with a “Webhooks by Zapier” GET step. Or run a cron that POSTs new rows to a Slack incoming webhook, your CRM’s REST API, or a Google Sheet. No proprietary connector required — it’s a GET request.
01Supply-chain / ESG risk manager at a large manufacturer
Enforcement Events feed →- Asks
- “Score my Tier-1 and Tier-2 suppliers by enforcement records across safety, environmental, and wage sources, trended year over year.”
- The feed returns
- Your supplier master matched by name and address to the feed: per-supplier counts of recorded actions, penalty totals, SNC/HPV flags, and back wages — with match confidence disclosed so you control the threshold.
- Action
- Reprioritize audits and dual-source where records are accumulating, before a continuity or reputational event.
02Corporate competitive-intelligence team
Workforce Filings feed →- Asks
- “Alert me when a competitor starts filing for roles in a capability we don't think they have — crypto, quant, LLM.”
- The feed returns
- First-appearance detection over job_title and SOC for named employers (punctuation-insensitive matching built into the employer parameter) — e.g. a regional bank's first-ever 'Quantitative Researcher - Digital Assets' filings.
- Action
- Brief strategy leadership on the capability build the filings document, before any product announcement.
03Freight broker / shipper — carrier vetting
Carrier Analytics feed →- Asks
- “Before I tender this load: what do the official records show for this USDOT number?”
- The feed returns
- One row: operating status, authority statuses on file, active insurance filings, safety rating and date, 24-month inspection, out-of-service, and crash counts, and the SaferGrade score — everything on record, nothing inferred.
- Action
- Apply your own tender criteria to a complete, current record instead of a stale lookup — and batch-vet an entire carrier list in one paginated pull.
For healthcare & med-tech sales
Healthcare & med-tech sales
For med-device, pharma, and diagnostics sales teams: the same Parse pattern over the CMS public files, served by our sister estate Healthparse (its own API key, $0.10 per call). Plain-English Parses, each a dated record change from a named CMS source file: a new provider enumerated or newly billing in your territory, a provider deactivated or reactivated, a hospital ownership change (CHOW), a reported capital-spending jump in a hospital’s own cost report, or a provider newly excluded or sanctioned. Healthparse supports webhook subscriptions on its enterprise tier — healthcare Parses can be push, not just pull (they are not delivered through ParseData saved Parses, which cover the ParseData feeds only).
Worked example
Build this Parse: new provider in my territory
For territory reps and commercial-ops teams at device, diagnostics, and pharma companies: every event is NPI-keyed and names the CMS source file it came from, so it joins your account master directly.
01First pull
Provider change events for your states — new NPI enumerations (NPPES weekly) and newly-billing providers (Order & Referring file, updated twice weekly). Drain pages with the cursor, same as every feed on this page. Hospital capex events live at/v1/hospitals/capex-signals, CCN-keyed from HCRIS Worksheet A-7.curl "https://api.healthparse.io/v1/providers/events?type=new_enumeration,newly_billing&state=TX,OK&since=2026-07-16&limit=1000" \ -H "x-api-key: ..."02Poll — or subscribe
Poll withsince=on your schedule and dedupe on each event’sdedupe_key— or register a webhook subscription and let Healthparse push new events to your endpoint.03Route on the record
Each event carriesevent_date,npi, entity name and address, and the source file plus vintage it was derived from — assign the new provider to the territory rep with the record attached.04Deliver
Anywhere HTTP reaches. In Clay, add an “HTTP API” column (save the key once as a headers account:x-api-key) and map response fields to columns. In Zapier, pair a Schedule trigger with a “Webhooks by Zapier” GET step. Or run a cron that POSTs new rows to a Slack incoming webhook, your CRM’s REST API, or a Google Sheet. No proprietary connector required — it’s a GET request.
Webhook push
Saved Parses API
Live for the ParseData feeds: save any composed Parse config with your API key and the gateway runs it on your cadence — daily at 12:00 UTC, or Mondays for weekly — and POSTs new rows to your webhook. The easiest path is the builder; the API underneath:
# save a trigger (returns id + signing secret — shown once)
curl -X POST "https://api.govparse.io/v1/triggers" \
-H "x-api-key: $GOVPARSE_API_KEY" -H "content-type: application/json" \
-d '{"name":"New TX carriers","webhook_url":"https://example.com/hooks/govparse",
"cadence":"daily","config":{"plain_english":"Every morning: new carriers in TX.",
"feeds":[{"event_type":"new_company","feed_id":"fmcsa.feed.carriers",
"params":{"state":"TX","added_since":"2026-07-22","limit":"500"}}]}}'
curl "https://api.govparse.io/v1/triggers" -H "x-api-key: $GOVPARSE_API_KEY" # list mine
curl -X POST "https://api.govparse.io/v1/triggers/<id>/test" -H "x-api-key: $GOVPARSE_API_KEY"
curl -X DELETE "https://api.govparse.io/v1/triggers/<id>" -H "x-api-key: $GOVPARSE_API_KEY"- Signed: every delivery carries
X-Govparse-Signature: sha256=<hex>— HMAC-SHA256 of the raw body with your Parse’s secret — plusX-Govparse-Delivery(uuid) andX-Govparse-Trigger(the Parse id). The body:{trigger_id, name, plain_english, new_rows, watermark}. - At-least-once: each run polls your feeds with the date filter set to the Parse’s per-feed watermark; the watermark advances only after your endpoint returns 2xx, so a failed delivery is re-sent next run. Dedupe on each feed’s stable key (
usdot,event_key,subject_key + last_marker_date). - Metered like a scheduled check: each executed page-call ledgers at the feed’s normal price ($0.05–$0.10) — a daily Parse costs the same as the daily check it replaces, empty runs included.
- Scope: ParseData feeds only. Healthparse legs in a saved config are marked
skippedin deliveries — webhook push for those is live via the Healthparse enterprise setup (hello@govparse.io).
These are examples, not SKUs.
Every feed is a generic, filterable API — the 30 recipes above are combinations of a feed, a filter, and a schedule that existing teams run, and any other combination works the same way. Machine-readable parameters, output shapes, and prices for every endpoint live in the catalog.