AlternativeSoft MCP · Model Context Protocol

Talk to your fund data. In any AI assistant.

AlternativeSoft MCP lets any AI assistant run AlternativeSoft analytics - directly, securely, in plain language. Bring AlternativeSoft into every AI workflow: the same trusted engine, the same data, with a natural-language front door.

Works with Claude ChatGPT Microsoft 365 Copilot GitHub Copilot Any MCP-ready AI
Claude Desktop · AlternativeSoft MCP
“Get me the funds in my sample hedge fund portfolio with their key attributes and stats.”
Used alternative-soft-mcp · resolved stat names · called  GetGroupAssets
8 funds returned - rendered as an interactive table:
FundStrategySharpeMax DD
FGC Multi StrategyMulti-Strategy1.88-16.7
ECM Enhanced GlobalMulti-Strategy1.70-14.9
NCM Absolute AlphaMulti-Strategy0.67-27.4
One Engine, Every Assistant
Plain-language analytics No menus, no exports Runs on your own data AI-agnostic & secure Server-Hosted or Local MCP Built for the AI era Plain-language analytics No menus, no exports Runs on your own data AI-agnostic & secure Server-Hosted or Local MCP Built for the AI era
The Brochure

A new AI-powered dashboarding experience

Get the full picture of AlternativeSoft MCP - how a question asked in plain language becomes live analytics, computed on your own data, and what that means for your team's workflow.

AlternativeSoft MCP - Brochure

The product, the use cases and the security model - in one PDF you can share with your team.

Download the Brochure

PDF · opens in a new tab.

The Shift

Your analysts are already talking to AI

AlternativeSoft MCP meets them there - turning natural-language questions into real analytics, run on your own data.

Built for the AI era

Future-proofs AlternativeSoft against the rapid shift to AI assistants and agentic workflows - without changing the engine your team already trusts.

Plain-language analytics

Ask for a Sharpe ratio, a peer comparison or a drawdown series - no menus, no exports. Just type the question and get a reasoned, data-backed answer.

How It Works

One protocol between AI and engine

MCP (Model Context Protocol) is an open standard that lets AI assistants call external tools. AlternativeSoft exposes its analytics as MCP tools - so the assistant can fetch funds, build comparisons and compute statistics on demand.

You

Ask in plain language

“Compare these three funds on Sharpe and max drawdown.”

AlternativeSoft MCP

Translates & orchestrates

Maps the request to the right analytics tools and parameters.

AlternativeSoft Engine

Computes on your data

Runs the calculation on your database and returns the result.

AI-Agnostic

Works with the AI tools you already use

No lock-in. AlternativeSoft MCP plugs into any assistant that supports MCP servers - and more are being added across the industry.

Claude Desktop

Claude Web

Claude Code

ChatGPT

OpenAI Codex

Microsoft 365 Copilot

GitHub Copilot

Any MCP-ready AI

Many Ways to Deploy

Server-Hosted MCP or Local MCP

Choose the deployment that fits each assistant. Web-based assistants consume Server-Hosted MCP; desktop and developer tools can make use of the MCP locally.

Server-Hosted MCP

AlternativeSoft Web API installation required

Supports browser-based assistants, desktop and developer tools that connect to a hosted endpoint.

Supported Assistants
  • Microsoft 365 Copilot
  • Claude Desktop
  • ChatGPT (Web)
  • Claude Code
  • Claude (Web)
  • ChatGPT Desktop
  • GitHub Copilot
  • OpenAI Codex

Local MCP

Runs on the user’s machine

Suitable for desktop and developer tools that launch locally - ideal for AlternativeSoft desktop users.

Supported Assistants
  • Claude Desktop
  • Claude Code
  • ChatGPT Desktop
  • GitHub Copilot
  • OpenAI Codex
Use Cases

What your team can ask for

From fund discovery to due-diligence prep - the same questions your analysts ask today, answered live and in plain English.

Fund discovery

“Find funds in my Long/Short peer group with the best 3-year Sharpe.”

Side-by-side comparison

“Compare Fund A, B and C on return, volatility and max drawdown.”

Statistics on demand

“Give me the rolling 12-month beta of this fund versus its benchmark.”

Portfolio insight

“Which funds sit in my Growth portfolio, and what are their weights?”

Screening & membership

“Does this fund belong to my Approved-Managers group?”

Due-diligence prep

“Summarise key risk stats for these managers ahead of the IC meeting.”

See It In Action

From a question to a live answer

Real sessions in Claude Desktop. The analyst asks in plain language; AlternativeSoft computes over MCP and renders the result.

A question becomes a live table

  • The prompt

    “Get me the list of funds in my sample hedge fund portfolio and some attributes and stats for them.”

  • MCP tool call

    Claude loads the AlternativeSoft tools, resolves the correct statistic names, then calls GetGroupAssets on your engine.

  • Live result

    Eight funds returned with strategy, currency, AUM, return, volatility, Sharpe, Sortino and max drawdown - as an interactive table.

  • Reasoned read-out

    Claude highlights the best risk-adjusted names and the highest-volatility outliers - analysis, not just data.

Claude · AlternativeSoft MCP
Get me the list of funds in my sample hedge fund portfolio and some attributes and stats for them.
Used alternative-soft-mcp · called GetGroupAssets
8 funds · stats as of Dec 2025
FundStrategySharpeSortinoMax DD
FGC Multi StrategyMulti-Strategy1.884.39-16.7
ECM Enhanced GlobalMulti-Strategy1.704.13-14.9
NCM Absolute AlphaMulti-Strategy0.671.23-27.4
ValueTech High ReturnDirectional1.142.73-18.5
Best risk-adjusted: FGC Multi Strategy leads on Sharpe (1.88) and Sortino (4.39), with a clean 8.2% volatility profile.
Claude · AlternativeSoft MCP
Pull Fundsmith Equity’s latest performance and risk stats and write a two-paragraph assessment I can use in the quarterly review.
Searched fund · pulled supported stats (GBP Acc)
Fundsmith Equity I GBP Acc - Q2 2026 Performance & Risk Assessment

Performance. A structured narrative grounded in returns, volatility and Sharpe across 1/3/5-year windows - built on the figures just pulled, with no invented numbers and no manual copy-paste.

Live stats become a write-up

  • The prompt

    “Pull Fundsmith Equity’s latest performance and risk stats and write a two-paragraph assessment I can use in the quarterly review.”

  • Pulls the real numbers

    Claude fetches the supported statistics for the GBP Acc share class - returns, volatility and Sharpe across 1/3/5-year windows - directly from your data.

  • Writes the assessment

    A structured Performance and Risk narrative, grounded in the figures it just pulled - report-ready prose alongside live data.

Best Practices

Get better answers with fewer prompts

A few habits keep responses fast, accurate and economical on tokens.

Be specific up front

Name the exact funds, group, statistics and date window in one prompt. Specifics avoid back-and-forth and wasted calls.

Mind token consumption

Request only the statistics and fields you need. Narrow scope returns smaller payloads - faster and cheaper.

Use groups & portfolios

Point the AI at an existing peer group or portfolio instead of long fund lists. One reference replaces many names.

Right-size the window

For time-series data, daily over decades is heavy; monthly is usually enough for comparisons.

Filter, then drill down

Start broad to shortlist, then ask follow-ups on the survivors - rather than pulling everything at once.

Confirm names early

Let the AI resolve exact asset names first. Correct names prevent failed lookups and retries.