Companion for Oh My Pi · community project

Stock Oh My Pi on a stock, local Strata server. Nothing forked, everything pinned.

Keep using Oh My Pi in your terminal. OMP Strata pins and verifies every download, installs Strata with its own unmodified setup.py from verified local inputs, starts it on loopback behind an API key, and launches stock OMP in an isolated profile where every model role points at that server. One Windows 11 + RTX host, one GPU you own, no proxy, no cloud fallback.

Install. From a checkout of the repository, in PowerShell; the quickstart has the exact tested commands. The integration lives in its own root; your normal OMP configuration and %APPDATA%\Strata are never touched.

py -3 scripts\omp_strata.py validate   --profile $prof
py -3 scripts\omp_strata.py fetch      --profile $prof --root $root
py -3 scripts\omp_strata.py install    --profile $prof --root $root
py -3 scripts\omp_strata.py keygen     --profile $prof --root $root
py -3 scripts\omp_strata.py start      --profile $prof --root $root
py -3 scripts\omp_strata.py launch-omp --profile $prof --root $root

Get started Read the decision

What the tool does

A profile under profiles/ is the single source of truth: host expectations, stock setup choices, and every artifact pinned by URL, size and SHA-256. A changed component or setting is a new profile with its own root and ledger.
fetch downloads the pinned OMP binary, the Strata engine and llama.cpp archives and the model shards, resumable, and verifies every byte before anything is used.
install runs stock Strata's own setup.py from those local inputs only, with its data folder redirected into the integration root, and checks the generated engine flags against the profile.
start binds Strata to 127.0.0.1 behind a user-only key file, refuses to start if the port, GPU or RAM budget is not free, and returns once authenticated identity checks pass.
launch-omp runs the pinned stock OMP binary in an isolated profile: every model role on the local server, discovery off, retries and model fallback off, ambient provider credentials scrubbed.
Gate probes, a tracer and a frozen six-task evaluation run on the real host; scrubbed receipts under releases/ bind each result to the exact profile fingerprint.

Use it when

You run Oh My Pi on Windows 11 with an RTX 5090, or an RTX 3090 or 4090 on a draft profile, and you want the model prompt to stay on hardware you own.
You want both stock projects exactly as released, every byte verified, with a rollback root for the previous candidate instead of an in-place update.
You want to read what was measured before you install: gates, throughput, context limits, restart cost and the evaluation, each tied to a receipt.

Use something else when

You want Strata's own app and installer: run stock Strata directly. This tool exists for the pinned, isolated OMP route.
You need session state that survives an engine restart. Here an engine restart re-prefills the whole transcript; durable continuation is OMP NInfer's claim, not this one's.
You need images, remote clients, another operating system, several tenants, or a GPU outside the profiles in the repository. None of those is supported.

What stays on your machine

Inference needs no network. Strata listens on loopback only, every model, control and metrics route needs the API key, and the key lives in a user-only file that is never put on a command line, printed, logged or committed. Stock OMP runs with an isolated home and profile; the launcher points its HTTP proxy variables at a closed loopback port, so OMP's startup catalog fetch fails locally while the Strata route stays direct, and no retry, fallback or discovery can route a request anywhere but the local server. The host is treated as single-user: this is not a sandbox, a multi-tenant service or an egress firewall. Full model: security boundary · operations.

Status

Candidate, not qualified. The current candidate, win11-rtx5090-coder-iq1m-131k-strata0.1.30-omp18.4.6 (stock Strata v0.1.30, stock OMP 18.4.6, Qwen3.8-Flash-Next Coder IQ1_M at a 131,072-token context on one RTX 5090), passes every integration gate on real hardware except one: a tool call cut off mid-arguments is executed with truncated arguments, an upstream defect fixed in Strata v0.1.31 and on OMP's main branch but in no OMP release yet. The six-task evaluation verified 15 of 18 scored attempts. Draft profiles for the RTX 3090 and RTX 4090 pass every gate except that same one. The gate ledger, measurements and upstream findings hold the evidence.

The OMP family

OMP NInfer — durable local inference
OMP Session Gateway — every live session, one private mobile page
OMP Oracle — asynchronous web oracles
OMP Strata — this project