Every source you follow gets read, cleaned, embedded, and stored in a private hybrid search index. Your own published work goes in too. Query it from Claude or Cursor over MCP, or let Content Marketing Ideas use it automatically every time it writes an article for you.
We crawl your sitemap and index every published article. The system knows what you've already written so briefs and articles can reference it, link to it, or refresh it instead of duplicating it.
RSS feeds, YouTube channels, subreddits, podcasts, news keywords. Anything you add as a source is indexed (or transcribed) into the same searchable corpus.
Vector search (bge-m3 embeddings) plus full-text FTS5, blended into one ranking. Long structured documents are split into section embeddings so a query can land on the right paragraph, not just the right page.
No random web crawling. Your corpus only contains things you chose to follow. Delete a source or a single document any time from the dashboard.
Generate an API key on your account page and drop the Content Marketing Ideas MCP server into Claude Desktop, Claude Code, or Cursor. You get a rag_search tool that runs hybrid search over your private corpus, plus tools for listing documents, adding URLs, and managing sources. Ask "what have I already written about X?" and get answers grounded in your own work.
When you turn a brief into an article, the pipeline expands the topic into five to seven targeted queries (keyword, related questions, entities, Wikidata concepts), retrieves the best passages from your corpus, and feeds them to Claude as grounded context with stable source IDs. The draft gets written against your own knowledge, not the open internet, and every fact carries a provenance chain back to the source document.
Free tier includes corpus indexing for up to 5 sources. No card required.
Start free