deployment-fact standard

Where Standard runs

Your environment. Your tenancy. Your access rules. Never our cloud. Standard runs where you already keep the company — on-prem, hybrid, or your own cloud account. We recommend. Your CTO decides. The graph moves with you if you change models. Standing up the new serving environment still takes real work.

This node Where Standard runs bounded-by Only connected sources

Three models. You pick; we recommend. What is identical across all three: you own the brain, the build is a burst, models are swappable, the stewardship fee is the same, and moving between models is a supported path: the graph moves; bringing up and cutting over the new serving environment is still real work.

Model 1 is full on-prem, air-gap capable. Model 2 is hybrid — the default — premises plus a receipted hop into your own VPC. Model 3 is self-hosted models in your cloud tenancy, not a third-party AI API.

Performance commitments are capability-level: sessions, working context, recovery. We do not publish raw throughput.

We install Mnemma on ourselves first. The same gates we ask you to walk, we walk.

Claims

What this page stands behind.

#claim-your-env

The brain runs in your environment, under your access rules. Never our cloud.

Applies to
Standard
Evidence
We never hold your data — so we cannot lose what we never have.
Since
2026-08-20
#claim-three-models

Three models: full on-prem, hybrid (default), or your own cloud tenancy.

Applies to
Standard
Evidence
Hardware bands are planning figures until the paid recount. Cloud in Model 3 means your tenancy.
Since
2026-08-20
#claim-you-pick

You pick the model. We recommend.

Applies to
Standard
Evidence
Five questions on the security page pick the model on a kickoff call.
Since
2026-08-20
#claim-capability-slo

Performance commitments are capability-level, never raw throughput.

Applies to
Standard
Evidence
Concurrent sessions, first-token time at working contexts, recovery objectives — not tokens per second.
Since
2026-08-20
#claim-customer-zero

We install Mnemma on ourselves first. The same gates we ask you to walk, we walk.

Applies to
Standard
Evidence
In progress; this page updates as gates pass. Not a completed case study.
Since
2026-08-20

Machine-readable → index.json · /ai-info · llms.txt

browse The rest of the brain