Use SeekAlgo with an AI assistant through MCP

SeekAlgo is a crypto strategy backtesting and paper-trading workspace. Its Model Context Protocol (MCP) server lets a compatible AI assistant use SeekAlgo tools after you sign in. You can describe a trading idea to your assistant; the assistant can draft a SeekAlgo strategy script, ask SeekAlgo to validate it, and run a historical backtest. Review the script and results before using the strategy further.

The MCP connection, market-data tools, strategy validation, backtesting, and paper-bot inspection are available now through a compatible AI assistant.

Connect an MCP client

The production MCP endpoint documented for this app is:

https://www.seekalgo.com/api/mcp

Add it as a remote MCP server in a client that supports Streamable HTTP and OAuth discovery. The client should open SeekAlgo's sign-in page and complete the OAuth flow. Do not paste a personal access token into a shared prompt. Your client may use its own connection UI or an MCP bridge. If the endpoint does not respond, confirm the production origin with SeekAlgo before troubleshooting a client-specific configuration.

The MCP endpoint is an API, not a documentation page. The authorization metadata is available at /.well-known/oauth-protected-resource and /.well-known/oauth-authorization-server on the same origin.

What an assistant can do today

Task	SeekAlgo MCP tool
Search available market symbols	seekalgo_market_search_symbols
Get recent price bars	seekalgo_market_get_recent_bars
Check whether a drafted strategy is ready for the engine	seekalgo_strategy_validate
Run a strategy on recent data	seekalgo_backtest_run_recent
Compare symbols and input sets, including split-period summaries	seekalgo_backtest_run_matrix
List paper bots and inspect simulated trades	seekalgo_paper_list_bots, seekalgo_paper_get_trades
Compare a paper bot with a backtest	seekalgo_paper_compare_to_backtest

For example, ask your connected assistant: “Create a simple BTC moving-average crossover strategy for SeekAlgo. Show me the script, validate it, then backtest it on recent data. Explain the assumptions and the main risk metrics.” The AI assistant writes the draft; SeekAlgo's MCP tools supply market data and run the validation and backtest. A strategy that passes validation is not evidence that it will make money.

What happens after a backtest

In the SeekAlgo app, an eligible strategy can be deployed as a paper bot. Paper bots use live Binance market data and record simulated orders and fills; they do not place real-money orders. Historical backtests are simulations too, and their results depend on data, assumptions, fees, and market conditions. Read the paper-trading guide and strategy-authoring guide for the current workflow and constraints.

Common questions

Can I create a strategy without writing code?

A compatible AI assistant can draft the script from your description while connected to SeekAlgo MCP. You can inspect and test that script through SeekAlgo.

Does MCP place live exchange orders?

No. The current public workflow described here is backtesting and paper trading. Paper trades are simulated.

Does a backtest prove a strategy will perform well?

No. A backtest describes how a strategy behaved under historical data and assumptions. Review the rules, trade list, risk metrics, and later paper results before drawing conclusions.