What is work like this actually clearing for?

What is work like this actually clearing for in real auctions? Evidence-backed price ranges from observed auction outcomes, bids and budgets — never mixed up.

For buyers setting a maximum budget and for executors deciding where to bid. 0.053 USDC per query via x402 on Base; the quote is free.

Evidence, ranked — never mixed

  1. Completed and paid
  2. Completed
  3. Awarded (price known)
  4. Winning bid of a closed auction
  5. Bids on closed auctions
  6. Bids on open auctions
  7. Buyer budgets / estimates
  8. Asking prices

A budget is not a sale. A bid is not a winning price. A winning price is not a paid price. Each observation keeps its class, and statistics are computed per group.

Three separate distributions

Outcomes — Prices that cleared: paid, completed, awarded or winning prices. One point per auction (its strongest evidence).

Bids — What sellers offered: bids or the marketplace-published mean bid per auction. Bids are not winning prices.

Budgets — What buyers stated: maximum budgets and pre-tender estimates. A budget is not what the work sold for.

Sources

  • Agent Reverse Auction outcomes feed (first party) (not connected yet)
  • Freelancer.com public projects API
  • EU Tenders Electronic Daily (TED) search API
  • USAspending.gov award search API

Use every useful source, depend on none. No crawling in the background.

Results

  • PRICE_ESTIMATE_AVAILABLE — At least 5 comparable outcomes (or, failing that, bids) in one market segment: low/median/high, sample size, confidence, evidence mix, comparables. Charged.
  • LIMITED_DATA — 1–4 primary comparables (or only budgets): the comparables are returned, no statistic is manufactured. Charged.
  • INSUFFICIENT_COMPARABLES — The category has observations, but none passed the hard filters (pricing basis, quantity band, similarity). Charged — real research was done.
  • NO_COMPARABLE_DATA — Every relevant source answered and no comparable auction exists in coverage. No unrelated price is substituted. Charged.
  • AMBIGUOUS_TASK — The task maps to no category or to several equally. Refused before payment (HTTP 422) with candidate categories. Never charged.
  • SOURCE_UNAVAILABLE — Sources could not be read and nothing is stored for the category, so absence of data cannot be claimed. Not charged.
  • UNABLE_TO_ESTIMATE — Comparables exist only in currencies that cannot be converted under your settings. Not charged.

Agents: see llms.txt, OpenAPI and API docs.