Feature deep-dive
Cost Console
Last updated
Every article in Mentionwell ships with a token-level cost ledger. Per-stage, per-model, per-site. The CFO's question — 'what are we spending on AI tokens for the blog?' — has an auditable answer. No other tool in the AEO/content space exposes this; we make it the default.
What gets ledgered
- Per article: total input tokens, output tokens, model breakdown, USD cost, total stages.
- Per stage: research, synthesize, outline, draft, reflection, rewrite, metadata, FAQ, embed, images, render — each with its own row.
- Per model: every distinct model invocation, with input + output token counts.
- Per site: monthly rollup. What is each domain in the portfolio costing?
- Per day / 7d / 30d / all-time: aggregate trend, exportable to CSV.
Why we ship this
Vendor lock-in is built on opacity. Most AI tools won’t show you token spend because the markup matters. Mentionwell’s pricing is the engine subscription + an honest per-article cost — the cost console proves we’re not over-charging. It also lets you compare cost-per-article across plan tiers, model choices, and site complexities.
Admin console (super-admin)
For ZipLyne (the parent operating Mentionwell), the admin console rolls up costs across every tenant. Surfaced for transparency, not as a customer feature.