A primary-source review of MPP, Payment Auth, Tempo and HTTP 402 for the week ending October 11, 2026.
Last week was two new commercial services and a documentation default that changed which stablecoin integrators see first. This week is catalog curation on both sides of the ledger: two additions, ten removals, and a spec clarification underneath — no new protocol capability, but a directory that’s a little more honest about what’s actually reachable through it.
Apify joins as a prepaid, spend-capped token
mpp PR #1056 adds Apify to the service directory, with the implementing commit wiring in the payment flow: an agent pays once and receives a prepaid, spend-capped API token, minimum $1, valid for 14 days, covering web scraping, crawling, data extraction and automation.
That’s a different purchase shape than most MPP listings — not a per-call charge but a time-boxed, capped allowance an agent draws down against. It’s also, like every catalog addition, a listing rather than evidence of use: being reachable and documented isn’t the same as anyone having paid for a token yet. Worth testing next is the token lifecycle itself — whether an agent can actually obtain, spend against and exhaust a $1 minimum token the way the integration describes, and what happens at the 14-day boundary if spend remains.
Tempo’s own API becomes a first-party listing
mpp PR #1055 adds the official, indexed Tempo API at api.tempo.xyz to the service directory, with the commit giving agents directory access to balances, transfers and payment activity.
This is Tempo listing itself in its own protocol’s catalog — the chain MPP is built on top of becoming one more thing an MPP agent can query and pay for. That’s worth noting rather than reading too much into: it’s a first-party convenience, not a sign of third-party adoption. What’s worth verifying next is whether the listed endpoints actually return the balances and transfer data the entry claims, since a first-party listing should be the easiest one in the catalog to hold to its own description.
Ten Locus-backed namespaces delisted
mpp PR #1053, with its commit, removes ten Locus-backed Abstract namespaces that had gone unavailable: company enrichment, email reputation, exchange rates, holidays, IBAN validation, IP intelligence, phone intelligence, timezone, VAT and web scraping.
Ten listings disappearing in one PR is a bigger cut than the routine single-entry cleanups prior briefs have covered, and it says something about how much of a general-purpose service directory can go stale behind the scenes before anyone outside the maintainers notices. The useful thing to monitor here isn’t this PR — it’s whether the same backing namespace produces another unavailability wave later, which would say more about Locus’s reliability than about MPP’s.
Underneath: a fee clarification and a health-coverage fix
Two smaller changes sit below the catalog moves. mpp-specs PR #375 clarifies client-arranged fee requirements for Tempo, tightening spec language rather than changing behavior. mpp PR #1051 improves unpaid-service health coverage — the same health-check machinery last week’s brief covered landing, now getting broader test coverage. A separate docs sync, PR #1052, brings documentation current with mppx 0.13.1.
None of this is adoption news, and it isn’t trying to be. A spec clarified before it’s misread and a health check given wider coverage before it’s trusted are both the kind of maintenance that only shows up as an absence of problems later.
What I’m taking from this week
Nothing here changes what MPP can do — it changes what’s actually reachable through it. Apify and the Tempo API are two new, unverified-by-usage listings; ten dead ones came out. The fee clarification and the health-coverage expansion are both hardening work on things that already shipped. Next week’s honest question is whether Apify’s token flow or the Tempo API listing show any verifiable calls, and whether the newly expanded health coverage catches the next unavailable service before it needs its own delisting PR.
I built MPP Pulse to keep those layers separated instead of blurring them into one undifferentiated “things happened” feed. It’s an open-source AWS agent that collects primary evidence and drafts a cited weekly brief for human review.