Skip to content
MCP server

Investbrain

Investment discipline MCP server with RAG, behavior pattern detection and price reminders

Pricing
Open source
Platforms
Windows, Linux, macOS
Version
0.1.0
Listing updated
today

Similar MCP servers

About Investbrain

Investbrain is an open-source MCP server for finance and payments. It is available on Windows, Linux, macOS, CLI, API and self-hosted.

The README presents InvestBrain as a personal discipline aid for experienced investors rather than an advisory service. It stores the reasons behind each trade idea, links new ideas to similar past decisions, and tries to surface recurring behavior patterns such as chasing rallies or panic selling.

The server exposes its tools over stdio to any MCP client, and the README also describes a Windows desktop build. Retrieval draws on a knowledge base of 16 investor frameworks and five core concepts, plus Chinese A-share market data from AKShare and Tushare. Reminders can fire on price or condition thresholds and push notifications to configured channels such as PushPlus, Feishu, DingTalk or Bark.

The project is aimed at investors who want to check their own decisions against their stated rules, and at developers who want to run a self-hosted investment journal with an LLM. Users supply their own DeepSeek key and any optional data-provider tokens; the README says these are kept out of git.

Tools this server provides (6)

Taken from the project documentation; the running server may differ.

thought_record_thought
Records a thought for an investment idea.
rag_ask_investment
Asks a question answered from the investor knowledge base.
memory_get_behavior_patterns
Returns detected behavior patterns from recorded decisions.
reminder_set_reminder
Creates a condition-based reminder.
market_get_stock_quote
Returns a stock quote.
notify_configure_notifier
Configures a notification channel.

What it does

Thought capture and decision cards
Records a short investment idea in plain language and turns it into a card linked to related past decisions.
Investor framework retrieval
Answers questions by searching a knowledge base of 16 investors and five core investment concepts.
Behavior pattern detection
Scans recorded decisions for recurring biases and reports the patterns it finds.
Price and condition reminders
Sets threshold-based reminders that trigger a review and send notifications to configured channels.
A-share market data
Provides quotes, price history, financial statements and valuation data for Chinese stocks via AKShare and Tushare.
Local data storage
Keeps memory and configuration in local SQLite and vector files rather than uploading them to a cloud service.

Good for

  • Reviewing a trade idea against past decisions before acting on it
  • Spotting repeated behavior errors such as buying after large gains
  • Setting price alerts that prompt a re-evaluation of a holding
  • Querying Chinese A-share financials and quotes from inside an MCP client

Pricing

Open source

AGPL v3 open source; cloud report and team plans are listed as planned.

Investbrain FAQ

Does InvestBrain give investment advice?

No. The README says it is not an advisory service and describes itself as a mirror and discipline anchor for the user's own decisions.

Which AI models does it need?

It requires a DeepSeek API key for the MCP server. DashScope is listed as an optional alternative LLM.

Where is data stored?

The README says memory and cards are kept in local files and SQLite databases, and the desktop client stores data under the AppData folder on Windows.

What does the paid tier include?

The README lists cloud weekly and monthly reports at 39 RMB per month and a team plan at 199 RMB per month as planned, not yet available.

Is Investbrain free?

Yes. Investbrain is open source, so you can use and self-host it for free.

What are the best alternatives to Investbrain?

The closest alternatives to Investbrain are Investboard, EchoPai (回声投研), jizhang, invest-like: Buffett-style stock analysis and hhxg A-share Market Data.

Where this comes from

  • Description: written from the project's README and its MCP Registry entry.

How AIPoweredBot checks servers