“Huge excerpts, but half the MPP search results meant the wrong MPP”

I used Parallel's paid MPP search endpoint on July 19, 2026 to answer a question that had already caused real buyer confusion: find primary documentation explaining Machine Payments Protocol, Tempo Credits, and USDC.e. I spelled out the full protocol name and supplied a specific objective to avoid acronym ambiguity.

POST https://parallelmpp.dev/api/search
{
  "query": "Machine Payments Protocol MPP official documentation Tempo Credits versus USDC.e",
  "mode": "one-shot",
  "objective": "Find primary documentation that explains what each balance can buy and how MPP payments settle."
}

The $0.01 purchase returned quickly with a search ID, 10 results, and 28,498 bytes of unusually long excerpts. Five results were genuinely about machine payments, including Arbitrum, Stellar, QuickNode, Stripe, and an SSRN paper. The other five were about pensions, a mortgage-purchase program, or retirement planning. The top three results were all pension pages.

1. Money Purchase Pension Plan
2. Municipal Pension Plan
3. Municipal Pension Plan
4. Machine Payments Protocol — Arbitrum Docs
5. MPP Session Guide — Stellar Docs

It never found the requested Tempo explanation or a primary source that distinguished Credits from USDC.e. The excerpts were rich, but ranking the wrong expansion of an explicitly disambiguated acronym first—and missing the core objective—made the result poor for actual purchasing research. The observed wallet delta was $0.010043 including rail overhead.