The Mentova assistant is a conversational AI co-pilot built into the dashboard. It reads the page you're on and helps you act on it — understanding an opportunity, sizing up a source, drafting content, or making sense of a prompt or competitor — all in a chat, without breaking your flow.

It's available on every plan and metered by your Agent credits pool (a modest allowance on Starter, generous on Pro and above; none after the trial). The deterministic analysis pipeline is unchanged: the assistant sits on top of it and helps you interpret and act, it never alters your metrics.


Where you'll find it

Mentova is available on every page of a brand (bubble in the bottom-right corner on desktop, "Mentova" tab on mobile). It always knows which page you're on — the panel header shows it ("On: Sources") — and it's especially useful here:

SurfaceWhat it helps with
Dashboard and AlertsExplain an alert in numbers and dig in: which queries one AI assistant (ChatGPT, Gemini, Perplexity…) does not cite you on, which brands it names instead and which sites it actually read, then propose the content to write
OpportunitiesExplain a gap, prioritise it, and draft the content to close it
SourcesAssess a source's influence and how to earn a mention there
ContentDraft, refine and adapt content for a specific gap
ConnectorsAdvise where to publish what, based on your connectors and publication history (connectors and publication history live on the same page, Automation › Publications)
PromptsInterpret why a prompt does (or doesn't) surface your brand
CompetitorsCompare a competitor's positioning against yours
SchedulingAdvise on a schedule frequency, models and prompt count that fit your plan and credit budget, and propose turning them on or adjusting them via a card you confirm

Conversations, history and memory

  • A conversation follows you from page to page. Start on the Dashboard and carry on on a specific gap: the panel stays open while you navigate, and Mentova knows where you are with every message.
  • A fresh thread after 6 hours without a message. When you come back, Mentova opens on what changed since your last visit — completed analyses, new gaps, alerts, pending drafts — with the real numbers and a link to each page. This card costs no credits.
  • History (the panel's ⋯ menu) lists your past conversations on the brand: reopen one to continue it, or delete it. Your conversations are yours: your team members have their own.
  • Brand memory (⋯ › Memory) keeps up to 20 durable facts — "we mostly publish on LinkedIn", "don't track X". Mentova remembers nothing without your approval: it offers a "Remember: …" card, and you can erase a fact at any time. This memory is shared by the brand's team. If a fact contradicts a measurement, the measurement wins.

Actions, with your approval

Mentova changes nothing without you. Every proposal arrives as a card: you confirm it, or you dismiss it.

  • Free and reversible — archive or restore prompts, rename a theme, hide a competitor, snooze an alert, mark a gap as done. Several cards of this kind can be accepted at once ("Accept all"). Exception: depositing an existing draft on your CMS (WordPress, Webflow, GitHub) is free too, but that card is always confirmed on its own — the moment a click touches a system outside Mentova, it earns its own confirmation.
  • Paid — write a draft (5 agent credits), rewrite it (1), produce a PDF (1, paused for now), run Automate (5, or 15 with an illustration), relaunch an analysis (research credits). The cost is on the card, and each one is confirmed on its own.
  • Recurring cost — turn on or step up scheduled analyses, measure one more assistant: nothing is spent right away, but the card shows the effect on your month ("≈ 60 → 120 research credits / month"). Lowering or turning off is free.
  • Irreversible — stop tracking a competitor, regenerate a theme's prompts, archive a whole theme: the card says what will be lost.

What Mentova never does from the chat: put content live, schedule a publication, send to Zapier or a social network. It can deposit a draft on your WordPress, Webflow or GitHub; going live stays a click you make in Publications.

Adding or removing a measured assistant is reserved to brand admins.


Multi-model, automatically

You don't pick a model — Mentova routes each message to the right one behind the scenes:

  • Simple questions → a fast, cheap model answers directly.
  • Anything that needs fresh, live information → routed to a web-search model that browses before answering.
  • Content creation and rewriting → routed to a strong writing model.

You'll see a small badge and a live status ("Searching the web…", "Writing content…") so you always know what it's doing.


What it can do

  • Answer questions in context about your visibility, competitors, sources and prompts.
  • Research the web to pull in current, sourced information.
  • Propose content generation — the same 8 formats as the Content Engine, on a card you confirm, then saved to your content history.
  • Propose a PDF of what it generated, on a card you confirm, ready to share (paused for now).

It stays on-topic: Mentova only helps with your AI visibility (your brand, its competitors, sources, tracked prompts and content). Off-topic requests are politely declined and never charged.


Agent credits

Every action draws from your monthly Agent credits pool — the same pool that powers content generation, and separate from your campaign/research credits. Costs are weighted by action:

ActionAgent credits
Chat message1
Chat message with web search2
Generate a piece of content5
Apply a fix (targeted rewrite)1
Re-evaluate a hand-edited draft1
Repurpose a piece into another format5
Illustrate a draft (generated image)10
Generate a PDF (paused)1

Monthly allowance by plan: Starter 75 · Pro 250 · Pro+ 1000 · Agency 2500. When the balance hits 0 the assistant is locked until you upgrade or top up — your campaigns and research credits are unaffected.

Tip: Ask Mentova right where the insight lives. Opening it from a specific opportunity or source gives it the exact context to give you a sharper, more actionable answer than a generic question would.