Back to resources
Beta · Plus

MCP Connector

Let Claude, ChatGPT, Cursor, and other AI assistants access your AcaTrove research record — projects, papers, grants, deadlines, and lab context — as structured memory through the Model Context Protocol.

Why use the MCP connector?

Most AI tools start every conversation from scratch. The AcaTrove MCP connector gives your AI assistant access to your actual research record, so it can answer questions, draft materials, and find connections using the papers, projects, grants, and lab data you have already organized.

Ask your AI assistant about your active projects and deadlines
Draft grant sections, bios, or summaries grounded in your real records
Search your documents and publications without leaving the AI chat
Check lab equipment availability or sample lineage conversationally

Available tools

The MCP server exposes 21 tools that AI assistants can call. Tools require scoped API tokens — your AI assistant only sees data you have authorized.

Documents & Publications

list_documentsList documents in your workspace
search_documentsSearch across documents by keyword or topic
get_document_summaryGet an AI-generated summary of a document
chat_with_documentAsk questions about a specific document
list_publicationsList publications in your workspace

Projects & Tasks

list_projectsList your active projects
get_project_statusGet structured context for a project
create_taskCreate a task on a project
create_noteSave a note or insight from an AI conversation back to a project
link_source_to_projectLink a publication to a project

Grants

search_grantsSearch grant opportunities across NIH, NSF, and EU CORDIS

Lab Operations

list_lab_samplesList lab samples with metadata
get_sample_lineageTrace a sample's lineage and history
list_lab_resourcesList lab equipment and resources
check_resource_availabilityCheck if a resource is available
list_protocol_runsList recent protocol runs
summarize_lab_notebookSummarize lab notebook entries

Workspace

list_deadlinesList upcoming deadlines across all projects and grants
search_across_workspaceSearch documents, publications, and projects in one query
draft_from_contextGenerate a bio or research summary from your profile and project context

Getting started

  1. Upgrade to a Plus plan or eligible shared workspace.
  2. Create a personal API token in AcaTrove Settings with the scopes you need.
  3. Install and configure the MCP server in your AI client (Claude Desktop, Claude Code, or any stdio-capable MCP client).
  4. Start asking your AI assistant about your research.

Detailed setup instructions and configuration examples are included in the MCP server package. The connector is currently in private beta while we harden packaging, documentation, and client compatibility.

Available in Plus and above

The MCP connector is available as a beta for Plus subscribers and eligible shared workspaces. Free workspaces do not include MCP access.

See pricingBack to resources