The Content Engine generates publication-ready content targeted at the specific gaps where your brand is missing from AI responses. Instead of writing from scratch, you feed it an opportunity or source gap context and it produces format-matched content you can publish or adapt.

Content generation is available on every plan, metered by your shared Agent-credits pool — see below. It's only locked when your credit balance hits 0 (upgrade or top up to keep generating).

Content formats

Eight formats are supported, each optimised for a different publishing context:

When you arrive from an Opportunity or a Sources gap, the format is pre-selected based on the source type detected.

How to generate content

From an opportunity

1. Open any opportunity on the Opportunities page.
2. Click Open the studio - the Content Studio opens, the Mentova panel receives the opportunity's context (topic, missing sources, competitors mentioned) and starts right away.
3. Mentova recommends a format and an angle, then generates a scored draft after your confirmation.

From a source gap

1. On the Sources page, find a gap domain and click Create content.
2. From the Pro plan up, the manual options open with the source type and domain pre-filled — click Generate.
3. On Starter (no manual options), the gap's context pre-fills the Mentova composer: press Enter to send it, and Mentova recommends a format then generates.

Standalone

Navigate directly to Content in the sidebar to generate content without a specific opportunity or source. Describe what you want to Mentova in the chat — it recommends a format and generates on confirmation.

From the Pro plan up, a "Manual options" section appears above the editor: fill in the context yourself, pick a format, and generate without going through the chat. On Starter, Mentova is the only door into the studio. A post-trial account (FREE) has no Agent credits left: the studio stays readable but generates nothing.

Reviewing and copying output

Generation runs asynchronously (usually 10-30 seconds). Once complete, the content appears in the editor area. You can:

• Copy as Markdown - ready to paste into your CMS or editor.
• Copy as HTML - useful for direct publishing.
• Expand / collapse the content in the history list.

Copying and exporting drafts requires the Pro plan or higher. On Starter, the draft stays readable in the studio but cannot be copied or exported — a notice on the studio's start screen says so before you generate.

Illustrating a draft

The Illustrate button, under the draft's title, generates an image from its content: an editorial illustration with no text (image models often write wrong words, so we never ask them for any), square for a post and landscape for an article. It costs 10 Agent credits, regeneration included. If generation fails, your credits are returned and the previous image is kept. Remove deletes it, free of charge. It is also available from Automate (the "Add an illustration" checkbox in the launch dialog, same price): the image is then created before the deposit and travels with the article.

The illustration travels with the article to your connectors:

• WordPress — it becomes the draft's featured image. The account behind your application password must be allowed to upload media (Author role or higher); otherwise the article is sent without its image, and the read-back says so.
• Webflow — fill in your collection's image field in the connector; without it, the article is sent without its image. The connector form is create-only: to add this field to an existing Webflow connector, recreate it.
• GitHub — the image file is added to the same Pull Request, next to the article, and referenced in the frontmatter under the image key (image: "./<slug>.png"). You can override that value by setting image in the connector's frontmatter; no other key receives the file path.
• Zapier — the image address is passed in the payload (imageUrl, empty without an image); your Zap decides what to do with it.

Content history

All generated pieces are saved in the Content page history panel. Each entry shows the format, title, and creation date. You can expand any entry to read the full content, or delete entries you no longer need.

Publishing to integrations

If you have integrations configured (PRO and above — connectors, manual publishing, going live and Automate move together; during the 14-day trial they are unlocked so you can watch the chain work on your own site), a Publish button appears on each content item. This sends the content to the connected platform (e.g. your CMS or API endpoint).

Whatever the path — manual publish, scheduled, or Automate — Mentova first asks your connector for a draft. On WordPress, Webflow and GitHub that draft is verifiable, and nothing goes live without an explicit gesture from you: the Publish live button once you have reviewed the draft, or the Write and publish live choice made when launching an Automate — a choice made per launch, never remembered (see Opportunities). Zapier is the exception: it has no verifiable notion of a draft, Mentova passes your choice to it in the payload (status, see below) and your Zap decides. Manual publications always ask for a draft — status: "draft" on Zapier — and never go live on their own; a scheduled publication is only possible towards WordPress, Webflow and GitHub: towards Zapier, where no draft can be read back and nobody will be at the screen at the scheduled time, Mentova refuses it (the publish menu says so as soon as a date is chosen) — publish now, with one click. Only an Automate launch sends status: "publish": the Write and publish live choice on a generic Zapier connector, or the single Write and publish to button on a social destination, where no draft exists (see "Social networks via Zapier" below).

Manage your connectors on the Publications page (sidebar, right after Content): it hosts a guided Zapier wizard for no-code automation, the classic WordPress / Webflow / GitHub / webhook configuration, and the history of everything you've published. The same connections panel is also reachable from the Content page toolbar.

Not every connector supports a verifiable draft deposit. WordPress, Webflow and GitHub can save a draft and then read it back straight from your site to confirm what it actually contains — on GitHub, that draft is a Pull Request, opened on a dedicated branch and never on your production branch. Zapier can do neither: it relays your content to your Zap with no confirmation that it landed and no way to read it back afterward. It remains a possible target for the Automate button (see Opportunities), but the launch dialog warns you before the click, and the arrival reads "Sent to", never "saved as draft": your choice between draft and live is passed to your Zap once, in the payload (status), and your Zap decides what to do with it.

Social networks via Zapier: LinkedIn, Facebook, X and Reddit

The Publications page carries a catalogue, "Zapier destinations": the five destinations Mentova can feed through Zapier are listed there, connected or not (WordPress, Webflow and GitHub are connected from the classic list right above), with the format it will receive and, for the networks, the note "Delivery can't be verified". The four networks go through Zapier — one connector per network, hence one Zap per network, whose trigger is "Webhooks by Zapier → Catch Hook" and whose action is "Create post" on the network in question. Clicking Connect on a row opens the Zapier wizard already pre-filled for that network; you still create the Zap on Zapier, paste the Catch Hook URL, then, once the sample content is received, map the action to its fields and publish the Zap. A fifth row, Other service, is the generic Zapier connector: Mentova does not know what its Zap feeds, so it imposes no format on it.

What Mentova sends, and what it does not know:

• The format follows the destination — you do not pick it: LinkedIn gets a LinkedIn post, X a thread, Reddit a forum answer, Facebook a Facebook post. That is what guarantees a blog article never goes out to X. From an Automate, that format is written directly, and the launch dialog announces it before the 5 credits are spent. A long article on LinkedIn stays possible: generate it in the studio and publish it by hand.
• Your Zap receives two more fields: destination (linkedin, facebook, x or reddit) and destinationFormat (linkedin, facebook, x_thread or forum_answer), on top of status and the existing fields. The format field keeps its previous meaning — the format of the generated content, social_post or forum_post — so a Zap that filtered on it keeps working, and an "Other service" connector, or one created before this version, receives exactly what it used to.
• Paid Zapier, or free Make. The "Webhooks by Zapier" trigger is a paid Zapier feature (Pro plan). Make.com offers the equivalent, "Webhooks → Custom webhook", on its free plan: its address goes in the same field as Zapier's, and everything else works the same.
• "Create and test" sends a sample piece of content — a title, a text and imageUrl (empty), with type: "test" and status: "draft", in exactly the shape of a real delivery. That is what makes those fields show up in Zapier or Make so you can map them. While your Zap is a draft, nothing goes out; if it is already published, the sample will be relayed like a real post. A generic webhook (outside the Zapier wizard) still receives the minimal { type, timestamp } ping.
• Mentova does not know whether the post went out. Zapier acknowledges the delivery, nothing more — and it acknowledges it even when the action that follows fails on its side. So the screen reads "Sent to LinkedIn", never "published". If a post is missing, your Zap's history is where to look.
• There is no such thing as a draft on a social network: what goes out is public immediately, and nothing can be reviewed or withdrawn from Mentova. Launching an Automate toward a network says so under its single button (see Opportunities).
• Reddit: the wizard warns you when you connect it. Automated promotional content gets accounts and domains banned there, and every subreddit has its own rules — answer the question first.

Two connectors on the same network: the first one created takes the catalogue row; the second stays in the connectors list, where you can disable or delete it. No connector ever disappears from the screen while it exists.

Connecting a connector

You no longer have to dig identifiers out of another dashboard: Mentova asks each platform for what it can list itself. Manual entry stays available on every screen, as the fallback — it is a convenience, never a required step.

WordPress — paste your site's address and click Connect my site. Mentova first checks that your site accepts the connection, then sends you to WordPress's own authorization screen, where you approve access in one click; you come back exactly where you left, connector created. There is no application password to create or copy: WordPress generates it and Mentova receives it. That password is revocable at any time from your WordPress profile (Users → Profile → Application Passwords) — you hold the handle. Four situations block this path, and the screen names which one rather than sending you to a page that would turn you away:

• your site is not on HTTPS — WordPress disables application passwords on an insecure site;
• the address does not answer like a WordPress site (check it, or enter the credentials by hand);
• your WordPress does not advertise application passwords — a version older than 5.6, a security plugin that disables them, or a certificate handled by a proxy (Cloudflare and the like) that leaves WordPress believing it is on HTTP;
• your site answered, but the authorization address it advertises is unusable.

In those four cases, the Enter the credentials by hand fallback opens with the reason above the fields: site address, username, and an application password created from your WordPress profile. A fifth message, "Automatic connection is unavailable on this server", is not about your site: it is a missing configuration on Mentova's side, and manual entry still works.

Webflow — paste your API token and click Continue: Mentova lists your sites, then the collections of the chosen site; you pick from a filtering field, with no siteId or collectionId to look up. Changing the site resets the collection, so a collection can never belong to a site other than the one chosen.

GitHub — paste your token, then pick the repository and the branch from the same filtering list; the fields specific to your site (directory, extension, frontmatter — see below) are still typed in, since no API can guess them.

Two things to know about these lists. An empty list is a real result ("this account has no site"), and a failure is never shown as an empty list: if Webflow or GitHub reject the token, the screen tells you so as-is. And on a very large account the list may be truncated — the screen says so; the filter only searches what was loaded, so if what you are looking for does not appear, use manual entry rather than concluding the token has no access to it. The token is sent to our server at each step, to query Webflow or GitHub on your behalf — but it is never stored until you click Save: an abandoned step leaves nothing behind.

GitHub connector

For a site generated from a repository (Astro, Hugo, Next.js, Keystatic…), the GitHub connector saves each piece of content as a Pull Request: Mentova creates a branch, adds the file to it, then opens the PR against your base branch. Nothing touches your production branch until the PR is merged — Publish live merges it, and your site's deployment does the publishing. A refused merge (branch protection, required checks, insufficient token permissions) leaves the PR open: you merge it yourself, and the screen hands you its link.

What to provide when connecting:

• Access token — a fine-grained personal access token, with the Contents (read and write) and Pull requests (read and write) permissions, on this repository only. A token that is too broad is the most common way to get this configuration wrong. It is asked first: it is what lets Mentova list what follows.
• Repository — picked from the list of your repositories (or typed by hand: owner/name, or the full repository URL).
• Branch — the base branch for PRs, picked from the repository's branches; left empty, it is the repository's default branch.
• Directory — where to save the files, for example content/blog/en.
• Extension — .md, .mdx or .mdoc.
• Frontmatter — a JSON object of fields added verbatim to each file, on top of the title and date, the only fields Mentova fills in (the meta description is not one of them: copying the title into it would hurt your SEO, and a proper description is not yet routed to the file — if your site requires it, add it here). Every generator has its own schema: a missing field can break your site's build, an extra one is ignored. Example: {"category": "blog", "draft": false}. Only text, numbers and true/false are accepted.

Credit usage

Content generation shares a single monthly Agent credits pool with the Mentova assistant, separate from your campaign/research credits. Each action has a weighted cost: a chat message costs 1 Agent credit, a chat message with web search costs 2, generating a piece of content costs 5, applying a fix or re-evaluating a hand-edited draft costs 1, repurposing into another format costs 5, an illustration costs 10, and a PDF costs 1. Monthly Agent-credit allowances by plan: Starter 75, Pro 250, Pro+ 1000, Agency 2500 (a post-trial account: 0). The remaining quota is shown at the top of the Content page.

Tip: Start with high-priority opportunities and high-influence source gaps - content placed on authoritative domains has a disproportionate effect on AI citation rates.

Related

• Opportunities
• Sources
• Plans & credits