Cost & FinOps Monitor · Documentation
Overview
What Cost & FinOps Monitor does in the ecosystem, how it connects, and what is simulated.
← Cost & FinOps MonitorWhat it does
Cost & FinOps Monitor is the single pane for what the ecosystem's AI layer costs. Every metered app writes a row to the shared cost ledger; FinOps aggregates those rows into total spend, the Cloud-versus-OSS split, cost-per-run, and a breakdown you can pivot by app, stage, mode, or day.
Its headline number is the realized Cloud↔OSS saving: what the OSS-recorded runs would have cost at Cloud rates, expressed as a percentage versus an all-Cloud baseline — the dollar value of the dual-mode posture made concrete. CONFIGURE, which is owner-only, sets a daily Cloud budget cap; if any day's Cloud spend exceeds it, the Operate tab raises an over-budget alert.
How it connects
FinOps is a pure reader of the shared C8 cost ledger — the seed history plus every runtime metered row that the other apps write (#02 concierge, #03 support, #04 recommendation, #05 review, #06 content, #08 contract, #18 fraud, #19 photo, #23 BI, #33 KYC). It consumes the cost.logged events those apps emit, so when you run a metered stage elsewhere and reopen FinOps, the new spend is there. It reads the ledger but owns no mutable config beyond the in-session budget, and it makes no canonical writes.
What is simulated, and the honesty note
This app is fully deterministic — it only aggregates the ledger, so it has no metered model stage of its own. OSS runs are labelled as recorded on local hardware at roughly electricity-only cost; Cloud runs are priced per token. The budget engine is illustrative and labelled as wired to the real engine in Stage-2, where the same method surface maps to a real cost store with no UI change. The dataset is synthetic, with no real spend or PII.