Skip to content

Put your next decision to work

A practical guide to working with Mavori: choose how you reach the agent, give it a focused task, and review the evidence before you act.

Open web appStart with the guide

Choose your starting point

Mavori's hosted agent

Start a conversation in Mavori with no external AI client to configure. Research, strategy work, and portfolio context are available through the product's own agent.

Meet the agent

Your AI client, connected by MCP

Use Mavori's capabilities from a compatible AI client through Model Context Protocol (MCP). MCP is a paid capability; review your plan and authorize the client from Mavori's connection settings.

Explore MCP connections

Give each task a clear next step

These example requests show how to frame the work. Answers depend on the data, account access, and capabilities available to you.

  1. 01

    Make a research question specific

    “Compare these companies and show where the evidence is incomplete.”

    Name the instruments, time horizon, and decision you are working toward. Check the sources and data coverage before building a thesis around the answer.

    Data and coverage
  2. 02

    Turn an idea into a test

    “Define this strategy, then test its rules under realistic costs.”

    Keep the strategy version, test period, and assumptions together. Look at drawdown, trade count, and the shape of the results before deciding what to investigate next.

    Backtests and results
  3. 03

    Bring the decision back to your account

    “Evaluate this trade against my selected account and prepare a proposal.”

    Review the account, sizing, order details, and applicable Risk Controls. A trade proposal is a draft; the approval card is where you decide whether to proceed.

    Trade proposals and orders
  4. 04

    Review what actually happened

    “Help me review the outcome and record what I learned.”

    Use Trade Logs and Journal to connect the decision with its outcome. Keep the reason for a trade alongside the result so your next review has context.

    Trade Logs and Journal

Keep the limits in view

AI responses can be wrong. Historical results do not predict future returns. Paper trading, live trading, and strategy deployment have different requirements; review the supported account and action before proceeding.

Read about controls and privacy