The Marcora MCP server hands any MCP-compatible AI assistant authenticated, scoped access to your governed brand context and proven blueprints.
It has never read your positioning, your audience segments, your brand voice rules, or which templates have actually worked. So you spend the first ten minutes of every prompt re-explaining your company, and you still get content that sounds like everyone else’s.
“These tools are brilliant at generating. None of them know your business.”
The MCP server gives any MCP-compatible assistant authenticated, scoped access to your Context Hub and Blueprint library. Instead of coming to Marcora to work, Marcora shows up inside the tools you already use.
Brand Foundation, writing standards, targeting dimensions, curated collections, and proven templates.
And any MCP-compatible client. The AI pulls context, generates on-brand, and saves back, in one flow.
Real tools, grouped the way your work is.
Pull the brand foundation, writing standards, and targeting dimensions. Search the Reference Library and get cited results. Write new findings back as context items.
Browse and apply proven templates. Draft new ones from a conversation, then finalize them. Import from the community exchange.
Spin up a project to organize a launch or campaign, optionally with a generated brief, and keep the content that follows tied to it.
Generate on-brand content from a blueprint or an open canvas, or drive the in-document assistant. You get a link to open, review, and refine in Marcora.
Generate a trackable, optionally-expiring share link, or export to Word, without ever leaving the conversation.
Package multi-step work into reusable processes the AI builds and runs on demand or on a schedule. No rebuilding the same flow twice.
Only so much of your material fits in the model’s working window for one piece. Plain retrieval pools everything into one pile and takes the closest matches, so the moment you write something product-adjacent, your product docs dominate the mix and your brand voice gets squeezed out. Marcora treats context as a budgeting problem, not a search problem.
Marcora runs a separate lookup for each layer and merges them. A loud source in one lane literally cannot spend another lane’s budget, so adding a giant collection never shrinks what your reference library contributes.
Writing for a CFO, in Healthcare, at the Decision stage? Marcora guarantees the single best chunk for each before the rest of the budget fills by overall relevance, scored offline against your whole library so even a sparse audience surfaces.
get_relevant_context returns context that’s already assembled, weighted, and guaranteed, not a raw pile of matches.
Plain-language asks, in the tool you’re already in. Pick one to see what happens behind the scenes.
Connect Marcora to your account in any client, authorize once, and the AI gets access scoped to your permissions.
For automation, scripts, and agents with no person in the loop. Same toolset, headless.
Step-by-step guides for each platform. More are on the way.
Desktop, web, and Claude Code. Companion Skills included.
Connection guideCLI, desktop, and Cowork. One connection, right in your terminal.
Connection guideWeb, via Developer mode. Your brand context, one setup away.
Connection guideEditor and Agent, over MCP. Your context, right in your IDE.
Connection guidePoint any MCP-compatible tool at mcp.marcora.ai and connect.
The free tier includes MCP context access. Not a trial, not a teaser, the real thing.