Atlas Sbw870763cBack to TradingHome
Help CenterMCPCoinrule MCP Tools: What Can Your AI Assistant Do?

Coinrule MCP Tools: What Can Your AI Assistant Do?

Last updated July 17, 2026

Explore the Coinrule MCP tools available for strategies, portfolios, balances, baskets, templates, backtests, and trading management.

Coinrule MCP provides a collection of tools that compatible AI assistants can use on your behalf. These tools cover strategy monitoring and management, portfolio information, exchange accounts, templates, baskets, and backtests, with access controlled through Read and Write permissions.

You normally do not need to call these tools by name. You ask your assistant for what you need in plain language, and the assistant chooses an appropriate available tool.

For a broader introduction, start with  What Is Coinrule MCP?  or visit the  Coinrule MCP overview .

How do Coinrule MCP tools work?

Each tool is classified as either Read or Write.

Read tools require coinrule:read.

Write tools require coinrule:write, which is available when you connect an assistant using Read + Write access.

The assistant only sees tools allowed by its connection.

Write tools are hidden entirely from Read only connections.

See  Coinrule MCP Permissions and Security  for more information.

What tools can inspect your trading strategies?

Coinrule MCP provides several Read tools for understanding what your strategies are doing.

list_strategies

Lists your trading strategies or bots.

This is useful for broad questions such as asking which strategies you currently have.

get_strategy

Retrieves the full configuration and metadata for one strategy, including a dashboard link.

get_strategy_status

Provides a live strategy snapshot, including documented information such as open positions, PnL, paper balance, and trigger proximity.

get_strategy_activity

Retrieves the event log for an individual strategy, including signals, lifecycle events, and evaluations.

get_strategy_trades

Retrieves filled trades and closed positions for a strategy.

get_strategy_pnl

Retrieves a PnL time series that can be used for charting.

get_strategy_feed

Provides a combined activity feed across strategies.

Together, these tools support conversational monitoring workflows.

For example, rather than asking for a specific tool by name, you might ask:

Show me my running strategies and summarise how they are doing.

Your assistant can then select the relevant authorised tools.

What portfolio and account tools are available?

The MCP includes Read tools for balances, holdings, signals, and connected accounts.

get_portfolio_balances

Returns cash or quote balances on connected exchanges, together with strategy allocations.

get_portfolio_holdings

Returns asset holdings, total portfolio value, and allocation.

get_recent_signals

Returns recent buy and sell signals across your strategies.

list_exchange_accounts

Lists connected exchange and broker accounts, including the venue and account IDs needed when launching live.

These tools can support questions such as:

What are my current portfolio holdings?
How much available balance do I have?
What signals have my strategies recently generated?

Which tools create and manage trading strategies?

Strategy creation and control tools require Write access.

validate_strategy

This is documented as the first stage of a two-step launch process. It parses a prompt or script and generates tests.

create_strategy

The second step launches a validated strategy on a selected venue.

create_strategy_from_prompt

This provides a one-step route from a plain-English request to a validated, tested, and launched strategy.

update_strategy

Changes a strategy's configuration and redeploys it.

start_strategy

Resumes a stopped strategy.

stop_strategy

Stops or pauses a strategy and can optionally close open positions.

Because stopping a strategy with open positions can have different consequences depending on whether those positions are closed, make your intention clear when asking an assistant to stop or pause a strategy.

Write-enabled actions on a live connected venue can involve real funds.

Coinrule's guide to  MCP and AI trading agents  provides more context on reviewing strategy logic and using AI-assisted trading workflows carefully.

Can your AI assistant browse Coinrule strategy templates?

Yes.

The list_strategy_templates Read tool allows the assistant to browse pre-built scripts from the Script Library and filter them by level.

The documented template categories include examples such as RSI, DCA, TradingView, and baskets.

This means you can explore existing strategy structures before deciding whether to create something from scratch.

What Coinrule MCP basket tools are available?

Coinrule MCP exposes Read and Write tools for baskets.

list_baskets

Browses the curated basket catalogue and your custom baskets, with category filtering.

list_basket_assets

Searches the available asset universe when you are building a custom basket.

create_basket

Creates a custom basket with two to 20 assets and target weights.

This requires Write access.

launch_basket

Launches a curated or custom basket as a strategy.

The documentation states that a basket can be launched on one venue or across multiple venues.

For multi-venue baskets, Coinrule maps each exchange to the matching connected account.

If a required exchange account is missing, the launch may fail until the relevant exchange is connected.

What backtesting tools are available through Coinrule MCP?

The documented backtest tools include both Read and Write capabilities.

list_backtests

Lists saved backtest runs.

get_backtest

Retrieves a summary for an individual backtest.

get_backtest_export

Retrieves raw trades or export data for a backtest.

run_backtest

Runs a single historical simulation and requires Write access.

run_backtest_scenarios

Compares multiple parameter sets in one pass and also requires Write access.

Backtesting can help you explore how a strategy would have behaved historically, but historical results do not guarantee future performance.

How should you ask your AI assistant to use these tools?

You generally do not need to mention the underlying tool names.

Instead of:

Call get_strategy_status.

You can ask:

Show me the current status and open positions for my BTC strategy.

For actions where multiple interpretations are possible, give more detail.

Useful information can include:

  • Asset or trading pair.
  • Exchange, broker, or paper venue.
  • Paper or live execution.
  • Capital allocation.
  • Timeframe and indicators.
  • Entry and exit conditions.
  • Position sizing.
  • Stop-loss or take-profit rules.
  • Basket rebalancing frequency.

This helps the assistant select an appropriate Coinrule tool and reduces ambiguity.

For broader information about AI-assisted automation, explore Coinrule's resources on  AI trading  and  AI trading copilots .

The public  Coinrule Agentic Trading MCP listing  also provides an overview of the official MCP server.

Frequently asked questions

Do I need to know Coinrule MCP tool names?

No. You can normally ask in plain language and let the assistant choose the relevant authorised tool.

Why can't my assistant see a particular Write tool?

Check your connection permissions.

Write tools are hidden unless the assistant has been granted the coinrule:write scope.

Can an assistant stop a strategy without closing its position?

The documented stop_strategy tool can stop or pause a strategy and optionally close open positions.

Be explicit about the result you want.

Can Coinrule MCP work with multiple exchanges?

The tools reference documents connected exchange and broker accounts and multi-venue basket workflows when the required accounts are connected.

Source:  Coinrule MCP documentation 

Was this article helpful?