Content Factory MCP Server

Turn any text into study and publishing material: Q&A pairs, reports, quizzes, flashcards, mind maps, data tables, slide decks and podcast scripts.

No LLM and no API key: the engine is extractive and runs inside the Actor container, so the same input gives the same output. It reshapes your text; it does not rewrite, paraphrase, translate or change tone.

Run it two ways — paste your texts and click Start, or connect Claude Desktop (or any MCP client) to the Standby URL and call the tools from your chat.

Run it as a batch job

Fill in Source texts with one or more texts, pick the Content types you want, and start the run. Each text is transformed into every selected type, and every result is written to the dataset as its own row.

Input	Meaning
texts	One or more source texts. Each must be at least 40 characters.
contentTypes	Which transformations to run on each text.
language	Sets the labels the tool writes itself (question stems, headings, speaker names): en and ko. Any other value falls back to English labels; the result then shows "language": "en" and "requested_language" with the value you sent. Sentences taken from your text stay in the source language.
reportStyle	blog, briefing or study_guide — only used when report is selected.

Leave Source texts empty and the Actor starts as an MCP server instead, shutting down on its own after the idle timeout.

Connect from Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "content-factory": {
      "url": "https://ntriqpro--content-factory-mcp.apify.actor/mcp"
    }
  }
}

Restart Claude Desktop. The tools are available immediately, and every result a tool returns is also written to the run's dataset.

Tools

Tool	Description	Price
qa_generate	Q&A pairs, one per source sentence that carries a key term	$0.15
report	Blog post, briefing or study guide, as structured sections plus Markdown	$0.12
quiz	Fill-in-the-blank questions with distractors drawn from the text	$0.15
flashcards	Term / definition cards, graded by how often the term appears	$0.15
mindmap	Mermaid mindmap source, ready to paste into any Mermaid renderer	$0.08
data_table	Figures found in the text (amounts, percentages, counts, negative values) with their surrounding sentence, up to 200 rows (truncated and total_found when there are more); a text with no figures gets a key-term frequency table	$0.15
slide_deck	Title, body and closing slides with bullets and speaker notes	$0.20
podcast_script	Two-speaker script with an estimated duration	$0.18
list_voices	Voice catalogue	Free

Text-to-speech is not available

tts_generate is still listed so existing clients get a clear answer, but it returns an "unavailable" message and is never charged: the speech backend it used has been retired and there is no speech model in this Actor. If you need audio, use podcast_script — it returns a narration-ready script you can feed to any TTS service.

How it works

Everything runs inside the Actor container. The engine is extractive: it scores sentences and terms from your text and reshapes them, so every line of output traces back to a line of your input. Nothing is sent to an external AI service, and nothing is invented.

Because it is extractive rather than generative, results are deterministic — the same text produces the same output — and quality scales with the source. A few sentences give you a few results; an article gives you a full set.

Pricing

Pay per event, charged per result delivered.

Free plan: each run returns up to 25 billable results, then stops with a single notice row explaining the cap and finishes normally. Paid Apify plans receive the full result set in one run.

A call that produces no result — empty input, a text below the 40-character minimum, tts_generate, or a text from which nothing could be built (for example no quiz question because no subject repeats) — is not charged. In a batch run it leaves one notice row (noticeCode: generation-failed or empty-result); an MCP tool call gets the same explanation back as an error reply that says no charge was made, and an empty-result notice row is written to the dataset. A charged row always holds at least one item, and calls that built nothing do not count toward the free-plan 25.

Legal

No scraping and no external data sources. All output is derived from the text you provide.

---

🔗 Related Actors by ntriqpro

Extend this actor with the ntriqpro intelligence network:

• supply-chain-risk-mcp — MCP server for supply chain risk
• video-intelligence-mcp — MCP server for video intelligence
• document-intelligence-mcp — MCP server for document intelligence

⭐ Love it? Leave a Review

Your rating helps professionals discover this actor. Rate it here.

Disclaimer

GENERAL DISCLAIMER: Output is generated automatically and may contain errors, inaccuracies, or omissions. Do not rely on output for legal, medical, or financial decisions without independent professional review. We provide no warranty, express or implied, regarding the accuracy, completeness, or fitness for any particular purpose of the generated output.