Mmcp.market

Head to head

ONCE — Exactly-Once Reliability for AI Agent Actions vs Lockstep

Same numbers we show everywhere else, lined up. Ratings come from users, the grade comes from our scanner, adoption comes from public registries.

ONCE — Exactly-Once Reliability for AI Agent Actionsio.github.GSterlingPress/once
B
No reviews yet

Prevent duplicate AI-agent side effects with idempotency, verification, and durable receipts.

Full listing
Lockstepio.github.lockstep-team-agent/lockstep
A
No reviews yet

Decision memory for AI coding agents: captures decisions once, briefs every agent before it acts.

Full listing

The short version

Neither has been reviewed yet, so the comparison rests on the scan and the public usage numbers. The safety scan favours Lockstep (91/100 against 79). ONCE — Exactly-Once Reliability for AI Agent Actions is hosted and callable through the gateway; Lockstep runs on your own machine. Lockstep has noticeably more public adoption.

 ONCE — Exactly-Once Reliability for AI Agent ActionsLockstep
User ratingNo reviews yetNo reviews yet
Safety gradeB79/100A91/100
AdoptionGrowingGrowing
GitHub stars010
Downloads / week–175
Runs whereHosted (remote)Your machine (local)
Gateway-readyYesNo
Tools219
Write actions1—
Price per callFreeFree
Licence—Apache-2.0
Last release2026-09-132026-09-19
CategoryAI and agentsAI and agents

Other comparisons

ONCE — Exactly-Once Reliability for AI Agent Actions vs Agent-Native ChatONCE — Exactly-Once Reliability for AI Agent Actions vs MCPONCE — Exactly-Once Reliability for AI Agent Actions vs AuscaONCE — Exactly-Once Reliability for AI Agent Actions vs ProtectONCE — Exactly-Once Reliability for AI Agent Actions vs FormioAll AI and agents servers