# Publish mentor answers and videos on your website

An article page takes the answers your mentors have already given in MentorDrop, and the videos in your libraries, and writes them as articles on a page of your own website: a browsable index and one page per article. Your clients and any visitor can read them without signing in to MentorDrop, and search engines can find them. This guide explains what gets published, how to install it, what to expect from search engines and how to polish each article.

## Quick answer

- **It lives on your website.** You paste two lines into a page of your site and MentorDrop renders the index and every article there. There is no public page on MentorDrop.
- **One content source:** a session, a library or a Knowledge Room. Choosing it does not start writing.
- **MentorDrop groups by topic, you publish:** one article per topic (from several answers or videos, or a single one); you can select many topics and publish them at once. They are written a few at a time and appear as they finish.
- **Articles are text.** MentorDrop writes each article from what the mentors typed, the transcripts of their audio and those of your library videos. The recordings and videos themselves are never published: they stay private in your account.
- **MentorDrop never names who asked.** Not the client's name, not their account, not their voice, not their attachments. What a client *wrote* is content, and content is published: read it before it is public.
- **You decide, and you are responsible for what is published.** There is no prior review: any article can be edited or hidden whenever you want.

---

## Publish your first articles

1. Open **Studio › Articles › Sources**, press **Write articles** and choose a session with public answers, a library that transcribes or a Knowledge Room. Before you press **Start** you see how many answers and videos will be grouped and about how many credits it costs.
2. Let its answers and videos organize into topics. Open **Topics**, review the proposals and choose which articles to write; choosing a source alone writes nothing.
3. Read a finished article and edit or hide it if needed. Keep **Articles written each day** (in **Settings**) at 0 while you want to request all writing yourself.
4. In **Publications**, create the publication, add the source and set the website page where you will show it.
5. Open its **Share** window, copy the installation code to that page and check both the index and an individual article as a visitor.

The sections below explain source eligibility, automatic writing, installation and later changes.

## Sources and publications

**Studio › Articles** has two tabs:

- **Sources**: the work. Each session, library or Knowledge Room has at most one set of articles, which Studio names after it ("Articles from the session Launch"): its topics and the articles written from them. Its **Settings** hold an optional **category** (a short label the widget shows beside its articles), an optional **source line** (a whole sentence you write, such as "Based on the answers from Launchpad Mentoring", which the widget shows under each article's title and the API hands to your site to show as is) and an optional **end link** (a button text and an `https` address that close every article, after the origin line; with an optional heading and line, the widget shows it as a closing band). The category, the source line and the end link are the same in every publication that shows those articles.
- **Publications**: your website. A publication is a page of your site with a title, an identifier and an address, and it shows the articles of one or more sources together, newest first. A source can appear in several publications. Its new articles appear on their own in every publication that shows it.

For example, Launchpad Mentoring publishes the questions from its launch program and from its membership at `launchpadmentoring.example/questions`, and only the membership ones on another page. Deleting a publication deletes no article; deleting a source's articles deletes their topics too and takes them out of the publications that showed them, which stay. The session, library or room is not touched.

Each published **article** has:

- **A specific title** and **a contextual subtitle**. The original question remains in Studio; it is not in the public article nor in the API.
- **The written text**, drafted automatically from the mentors' text messages, the transcripts of their audio and those of the library videos, in the mentor's voice: the first person ("I") when one mentor answered the source's questions and taught its videos, the team's ("we") when several did; the title too. Who teaches the videos is set in **Taught by**, on the library or on each piece; with nobody chosen, they are taught by your team's only member, and with several members they count as one more voice. It is not the literal transcript: it is ordered and cleaned up. It is always written in your account's language, also when the answers are in another one: then it is translated faithfully, with figures, units and product names unchanged. Topics and the article's labels, such as the closing line, are in the account's language too. Each article is checked sentence by sentence against those answers and videos, from the title to the closing: what they do not back is corrected once and everything is checked again; what is still not backed leaves the article, and Studio shows the state and any reason that needs attention. When the title does not hold or what remains no longer answers the topic, the article is not published and the previous version stays. Unchecked text is never published: if the automatic drafting does not respond, the article waits and is retried on its own. The check can miss things, so read what you publish.
- **The mentors** whose answers or videos it uses, in the API only (`mentors`), for your website to show if it wants; the widget does not name them.

## Choose content: topics

| Source | What goes in | When to use it |
|---|---|---|
| **A session** | Its answered questions. | Working with content from one session. |
| **A library** | The chapters of its videos. | Writing from a course or recorded sessions. |
| **A Knowledge Room** | Answered questions from its sessions and the videos of its libraries. | Managing a set of sessions and libraries in one room. |

When you choose the source, MentorDrop **groups its answers and videos into topics**: those about the same thing in different words or about different cases end up together, with a proposed name. Grouping takes a few minutes on large sources and writes nothing. With **Group automatically** on (when you start, and in **Settings**) it keeps going on its own every few minutes with anything new; off, anything new waits and the **Topics** tab offers **Group now**. While answers or videos are still being organised the list says so; it never shows them as loose topics.

### Videos

A library video is written from chapter by chapter. To write from it, the library must transcribe (**Transcribe this library**, on its page): the transcript tells the speakers apart and is split into chapters. Whoever speaks the most in the video is the one teaching: what they say is what the article may state as the mentors' view. What others say, such as the students of a recorded session, only gives context and is never presented as said by the mentors; a chapter where only they speak gives no article. Videos are never linked publicly: in Studio, each one leads to its page.

A source's articles open directly on their **list of articles**; the **Topics** tab lists the proposals still without one. Every article shows one state in plain words — **Writing** (anything in progress: a first draft, a rewrite, the new questions being added, a wait for credits), **Ready** or **Failed** — or **Hidden article** when you took it off your website yourself; a short line under the state says why when it matters (waiting for credits, adding two new questions, a source question that stopped qualifying). On the article's page there is at most one main button: **Resume writing** when it failed, **Show** when it is hidden or **Publish** when it has not been written yet; a ready article needs none. You can search by article title and filter the list by **All articles**, **New questions** or **Hidden articles**; the numbers on each filter follow the search.

- **Write article**: one article that answers the topic, written from all its questions together. It opens with a short lead (one or two sentences that answer the question or lead into the article), continues in paragraphs, lists or subheadings, with one real case from the answers as an example of each idea that has one, right after it (the situation, with the age and sex the question gives but never a name, and what was decided), and may end with a short closing paragraph. On a ready topic, when you confirm you see roughly how many credits it will use. A topic marked **Not checked yet** has its answers and videos checked first.
- **Write many at once**: tick the topics and press **Write N articles**, or press **Write all (N)** to write every ready topic that matches your search (up to 500 at a time). They are written a few at a time, and Studio tells you roughly how many credits they will use: the sum of what each topic costs.
- **Articles written each day**, in **Settings**: how many ready topics are written on their own every day, those with the most questions first, as if you had asked for them; above 0, new questions are also added to published articles. It is 0 when you start, and 0 means only what you ask for is written. The article list says what the next articles to be written cost, worked out from the length of their answers and videos; **Write article** and **Write all (N)** use the same per-topic figures. An article with more material costs more. Without credits it waits, and a day that passes without articles is not made up.
- **View topic** opens its answers and videos. **View original answer** opens an answer in your account and **Open the video** a video; **Remove from topic** takes it out of this topic, and MentorDrop looks for another place for it.
- A topic built from a single answer or video is checked like any other: once ready it is listed with the rest of the topics.

A topic without enough material for an article of its own is shown greyed out, with the reason, and cannot be written.

A topic that is already published keeps gathering the new questions that resemble it. With **Articles written each day** at 0 they stay on the article's page until you press **Add N new questions now**. Above 0, MentorDrop first compares them with the article, and only when they add something the article does not say yet does it write the article again from all its questions: as soon as three such questions have gathered, or otherwise after a day without new ones. Meanwhile the article shows **Writing** with the number of questions being added, and your website keeps showing the previous version. On the article's page you can **Remove from topic** any question on its way in, and **Add N new questions now** skips the wait. An article you edited by hand is never rewritten on its own: only that button does it, after warning you that your text will be replaced (the title and link are kept).

Articles are written a few at a time; each one turns **Ready** and appears in your publications when it is done. You can cancel pending articles from the header menu; articles already written are kept. Choosing a source or making a publication visible never writes anything; only **Articles written each day** writes without you asking.

An article appears once it has a mentor answer or a video and its text is ready. Rejected questions and questions from clients who deleted their account are never published.

### Private questions

A private question is never published. Since every question is private in a session without the "Public answers" option, **to publish a session that option must be on**.

## What is never published

- The name of the client who asked, or any reference to their account: MentorDrop never adds them as a field, neither to the page nor to the API.
- Their voice recordings, their videos, their attachments and their form data as separate fields.
- The messages the client sent after the answer.
- Unanswered or rejected questions.

**What the client wrote is a different thing, and reviewing it is yours.** MentorDrop leaves out who asked, not what they said: a question that opens with "Hi, I'm Ana Ruiz", a case that describes a client by name or an attachment called `contract-ana-ruiz.pdf` is content, and content gets published. The article is written to keep the case anonymous, and neither the public template nor the API carries the original question, but MentorDrop does not filter what the client told on its own. Read what goes out — the article, its title and its address — and edit it whenever it names somebody.

## Install the page on your website

In **Studio › Articles › Publications**, open the publication and press **Share** at the top. Copy the snippet and paste it into an **HTML code block** on a page of your website dedicated only to this. You do it once. This is an example of its shape:

```html
<div id="mentordrop-content"></div>
<script src="https://your-brand.mentordrop.com/articles.js" data-mentordrop-page="questions"></script>
```

- The `div` is where the content is rendered. If your page already has another identifier, add `data-mentordrop-target="#your-identifier"` to the `script`.
- `data-mentordrop-page` is the publication's identifier, the same one you see in Studio.
- **The address of the page** where you paste it is the one you set in the publication's **Settings** (for example `https://launchpadmentoring.example/questions`). If you move the page, update it there.

It works on any page that accepts a script: your own website, WordPress, Kajabi and most platforms with HTML blocks. It uses no iframe: the index and the articles are part of your page.

The code only works on the domain of the page address you set in **Settings**, and on its subdomains (`www.launchpadmentoring.example` and `launchpadmentoring.example` count as the same site). Loaded from any other domain, the page shows a notice naming the expected domain and the one it was loaded from, and no article. A test or preview site must use that same domain, or you set its address in Settings while you test and put the real one back afterwards.

### Why a dedicated page

Every article has its own address within that page, in the form `?mentordrop=123-title-of-the-article`. The index is paginated with `?mentordrop_page=2`. The script sets the page title, description and canonical address according to what is being shown, and overwrites the ones your website editor may have set. That is why it must be dedicated to this page: if it shared space with other content, search engines would receive conflicting signals.

## Integrate your own website through the API

A website with its own design can read published articles as structured data. The widget remains available. Use your account host and page identifier:

```text
curl -H "X-MentorDrop-Key: YOUR_TOKEN" "https://your-brand.mentordrop.com/api/v1/publications/questions/articles?page=1&per_page=50"
curl -H "X-MentorDrop-Key: YOUR_TOKEN" https://your-brand.mentordrop.com/api/v1/publications/questions/articles/123
```

Every request carries your account's integration token in the `X-MentorDrop-Key` header (`Authorization: Bearer YOUR_TOKEN` also works). It is the same token that grants external systems access to your spaces: generate it or regenerate it in **Settings › Integrations**, where it is shown once; the publication's **Share** popup links there. Keep it on your server, never in a page's JavaScript. Regenerating the token stops every integration that used the old one.

The list returns `version`, `publication` (with `slug` and `title`), `data` and `meta` with `current_page`, `last_page`, `per_page` and `total`. `per_page` accepts up to 50; a page beyond the last one returns `404`. Articles arrive newest first by creation date, across every source of the publication; the order is not chosen.

Each summary carries, in this order, `id`, `slug` (the `{words}-{id}` key used in `?mentordrop=`), `title`, `subtitle` (may be empty), `description`, `mentors`, `word_count`, `updated_at`, `created_at`, `category`, `cta` and `origin_line`. `description` uses that subtitle when present, or an excerpt from the text. `mentors` are the mentors who answered or taught its videos, in article order, each with `name`, `initials` and `avatar_url`; an author who is no longer available is left out. Both dates are ISO 8601: `created_at` is when the article was created and the one to use for dates and archives, while `updated_at` moves with every edit or republish. An article's identifier is its own, not the question's. `category` is its source's category and `cta` its end link, with `label` (the button text), `url`, `heading` (the band's heading) and `line` (its sentence), those two null when you left them empty; `category` and `cta` are null when its source has none. `origin_line` is its source's **source line**, or null when empty: a complete sentence, shown as is next to the authors, with no words added around it.

The detail adds `text` (the whole article as plain text, closing included), `conclusion` (the closing paragraph) and `sections`. Each section has its `number`, `heading`, `key_answer` (the direct answer it opens with), `blocks` and the section `text`. Each block has `type`, `text` and `items`: a `paragraph` or a `subheading` carries its `text`, and a `list` carries its entries in `items` with an empty `text`. `sections` and `conclusion` come empty once you rewrote the text by hand: show `text` then. Text fields are plain text: escape them when rendering HTML.

The API is read-only. It follows the same publishing decisions as the widget, including your edits. A missing or wrong token returns `401` with `{"version": 1, "error": {"status": 401, "code": "unauthorized"}}`, also while the account has not generated a token yet. An unpublished article or an unknown page returns `404` with `{"version": 1, "error": {"status": 404, "code": "not_found"}}`; a page temporarily unavailable because of its plan returns `503` with a `Retry-After` header. Reading never requests writing or transcription.

In Astro or another website with a server, read the API when serving the page and insert the data into your template. You can then publish and edit from Studio without rebuilding the website. Your integration manages its own URLs, metadata and sitemap.

Successful responses carry `Cache-Control: private, max-age=0, must-revalidate`, an `ETag` and `Access-Control-Allow-Origin: *`. No shared cache or CDN stores them, since they depend on your token; your server can keep its own copy and revalidate it with `If-None-Match`, which answers `304` without a body when nothing changed. Errors are not cached.

## What to expect from search engines

- **Google and Bing** run the script and can index the index page and every article on your domain, with its title and description.
- **Crawlers of assistants such as ChatGPT and social media cards** do not run JavaScript: they see an empty page. If you share an article on social media, the card will show your page's data, not the article's.
- A custom API integration can deliver the article in your website's initial HTML. This requires implementing the server-side read; pasting the widget does not enable it.

### The sitemap

Each page that uses the pasted code has a sitemap with the addresses of its articles on your website. To hand it to Google, add this line to your website's `robots.txt`:

```text
Sitemap: https://your-brand.mentordrop.com/p/questions/sitemap.xml
```

WordPress and most hosting platforms let you edit that file, and the publication's **Share** popup shows the exact line. Kajabi does not allow it: there Google still discovers the articles by following the links the widget renders on your page. While the publication is not visible, the sitemap answers "not found".

A website that reads the API does not need this: it builds its own pages and its own sitemap.

## Polish an article

Click an article in the list to open it: the text as your visitors read it and the answers and videos it uses, each answer linked to its original question and each video to its own page. What the check took out is not listed: neither your answers nor your videos backed it. There you can:

- **Edit** the title and the article text. Changing the title does not change the saved link. MentorDrop does not rewrite what you edited on its own.
- **Hide** the article, and **Show** it again whenever you want.
- **Resume writing** when writing failed. While an article is being written again, and if the new writing fails, your website keeps showing the previous text; the new one replaces it only when it finishes.

From the header menu of a source's articles you can cancel its pending articles. Existing articles are kept; cancelling a regeneration preserves the previous text. Each session, library or room has one set of articles, started with **Write articles**. Those of a session or a library can be widened with **Widen to a room** to a room that holds it and has no articles of its own: everything else in the room joins and nothing is cancelled. Removing sessions or libraries from the room or videos from the library, or deleting the library, cancels pending articles that no longer belong to the source; articles already written are kept, and stay in the list even when their questions or videos are no longer in any topic.

## Timing

- Writing starts when you press Write article on a topic or a selection. Articles are written a few at a time and each one turns Ready when its text is done; there is no fixed completion time. If the mentor audio transcript is missing, that article waits until it becomes available. A video joins the topics once it is transcribed. With **Articles written each day** at 0, the scheduled process only advances what you already requested; above 0 it also writes that many topics every day.
- Hiding an article or editing its text takes up to **five minutes** to show for new visitors. Anyone who already had the page open sees it after reloading.

## Cost

Writing an article uses your plan's credits, at what drafting it really costs; before you write a ready topic or several at once, Studio shows roughly how many credits it will take. Articles are available on every plan and there is no limit on how many you publish.

## Troubleshooting

| Problem | What to do |
|---|---|
| The page is empty | Check that the editor keeps the `script` and the `div`, that the publication identifier is correct, that the publication is visible (its **Visible on your website** switch) and that it shows the article's source. |
| An article does not appear | Check that you wrote its topic's article, that it has a mentor answer, that its question is not private, that the article is not hidden and that its text is ready. Choosing a source does not start writing by itself. |
| The text is not written | Check that the session transcribes the mentors' audio and that the transcript has finished. If the mentor answered only by audio and transcription is off, there is no text to write. |
| A library gives no topics | Check that the library transcribes and that its videos are transcribed. If they are and grouping has finished, none of its videos teaches something to write an article about. |
| Google indexes the main page but not the articles | Use a dedicated page, add the `Sitemap:` line to your website's `robots.txt` and check that the address you set in Studio is exactly the page's address. |

## Related guides

- [Put MentorDrop on your platform](https://mentordrop.com/docs/embed-on-your-platform.md)
- [Client privacy and data](https://mentordrop.com/docs/participant-privacy.md)
- [How MentorDrop works](https://mentordrop.com/docs/how-it-works.md)

## How an article is laid out

Every article opens with a **short lead** (one or two sentences) and continues in paragraphs. When the mentor enumerates steps, options or products they appear as a list; a long article carries subheadings. MentorDrop decides that structure from the text: the article never contains HTML, and the look (typography, colours, spacing) is your site's.

Every article has its own page in the Studio: you see the text as your visitors read it and the answers and videos it uses, each with its link. From there you can hide it or edit the title and the text by hand; the topic's new questions are added with **Add N new questions now**, or on their own with **Articles written each day** above 0. While a new version is being written, your website keeps showing the previous one.