The Model Context Protocol (MCP) is an open standard that lets AI assistants call external tools. AlternativeSoft exposes its fund analytics as MCP tools - so Claude, ChatGPT or Copilot can screen managers, build peer comparisons and compute risk statistics on your own data, in plain language.
MCP standardises how an AI assistant discovers and calls tools. AlternativeSoft publishes its analytics as MCP tools, so the assistant can fetch funds, build groups and compute statistics on demand - then explain the result.
“Compare these managers on Sharpe and max drawdown.”
Maps the request to the right analytics tools and parameters.
Statistics are calculated on your data and returned as tables and prose.
Your team is already asking AI. MCP decides whether the answer comes from your governed data or from whatever the model can find.
Screening a peer group and pulling risk statistics used to mean menus and exports. With MCP it takes one plain-language request.
Numbers are calculated by the AlternativeSoft engine on your database - not estimated by the model - so figures are defensible.
The MCP runs against your own installation, server-hosted or local. Your holdings are never exposed to train anyone's model.
One integration works across Claude, ChatGPT and Copilot. Switch assistants as the market moves without re-platforming.
The risk in AI is not the question - it is where the answer comes from.
Each answer names the tool it called and the data it used, so a figure in an IC paper can be traced to the request behind it.
Holdings and returns are computed in place. Nothing is uploaded into a public model's training data.
The engine that already powers your reporting answers the AI - so a number in chat matches the number in your platform.
AI-agnostic by design - use the assistant your firm has approved and change it later.
Manager screening, peer comparison and write-ups collapse from an afternoon into a conversation.
Give teams a compliant way to use AI on fund data instead of pasting holdings into consumer tools.
Common questions about AlternativeSoft MCP and how it works with your AI assistant.
MCP stands for Model Context Protocol - an open standard that lets AI assistants securely call external tools and data sources. AlternativeSoft MCP exposes the platform's fund analytics as MCP tools.
Yes. AlternativeSoft MCP runs against your own AlternativeSoft installation - server-hosted or local - so statistics are computed on your data in place. Your holdings and portfolios are not exposed to train any AI model.
Any MCP-ready assistant, including Claude (Desktop, Web and Code), ChatGPT, Microsoft 365 Copilot, GitHub Copilot and OpenAI Codex.
No. The AI orchestrates the request, but every statistic is calculated by the AlternativeSoft engine on your database - there are no invented figures to re-check.
No. You ask questions in plain language inside your AI assistant. AlternativeSoft MCP handles the translation to analytics tools and parameters behind the scenes.
Tell us about your team and we'll walk you through AlternativeSoft MCP on your own portfolios and managers.
Prefer to read first?