Connect to Dafinchi MCP

Connect Dafinchi to your AI assistant to ask questions about companies, earnings calls, financials, and market data. MCP (Model Context Protocol) is how your assistant connects to Dafinchi. It can read information for your research, but cannot make changes to your account.

Connection settings

You need a Dafinchi account with MCP access enabled and an AI application that supports remote MCP connections with browser sign-in (OAuth). Dafinchi hosts the connection for you; there is nothing to install or run for Dafinchi itself.

MCP server URL

https://mcp.dafinchi.ai/mcp

Copy this MCP server URL exactly as shown into your AI application's MCP settings.

Setting	Value
Connection name	`dafinchi`
Connection type (if asked)	HTTP / Streamable HTTP
Sign-in method	OAuth — sign in to Dafinchi in your browser

Your AI application handles the connection details. Follow its browser sign-in prompts; you do not need to copy passwords, tokens, or API keys into the connection settings.

Connect from your AI application

Open your AI application's settings and look for MCP servers or connectors. Choose the option to add a remote server.
Name it dafinchi and enter the MCP server URL above.
Choose HTTP / Streamable HTTP and OAuth if the application asks.
Connect, then sign in to Dafinchi and approve access in the browser window that opens.
Return to your AI application and check that Dafinchi is connected. Try one of the questions below.

Your AI application keeps the connection signed in automatically. If it asks you to sign in again, follow the same browser steps. The tools available to you depend on your Dafinchi account's access.

Connect from the command line

Other AI agents, apps, and web interfaces that support remote MCP and OAuth use the same MCP server URL and sign-in process. Commands and menu names vary by application. Copy the MCP server URL from Connection settings above.

If you use Claude Code, open a terminal in your project folder and run:

claude mcp add --transport http dafinchi https://mcp.dafinchi.ai/mcp

Open Claude Code in the same project and type /mcp. Select dafinchi, follow the sign-in prompts, and approve access in your browser. Then return to Claude Code.

To check your saved connection settings, run:

claude mcp get dafinchi

For more help setting up Claude Code, see its official MCP guide.

If you use the Codex command-line app, open a terminal and run:

codex mcp add dafinchi --url https://mcp.dafinchi.ai/mcp

If you have not been prompted to sign in yet, run:

codex mcp login dafinchi

Sign in to Dafinchi and approve access in your browser. Then open Codex and type /mcp to check that Dafinchi is connected.

To check your saved connection settings, run:

codex mcp get dafinchi

For more help setting up Codex, see its official MCP guide.

Try it after connecting

Ask your assistant:

"Find Apple and show which earnings periods are available."
"What did Microsoft say about AI demand in its latest available earnings call?"
"Compare the latest fundamentals of Apple and Microsoft."
"Find published yearly reviews for semiconductor companies."

Your assistant chooses the tools for each question. Results depend on the information available in Dafinchi, so the latest earnings call available may be from an earlier quarter.

Research tools

With MCP access, your assistant can use these 16 tools to research companies for you. Ask a question in your own words; you do not need to learn the tool names or enter technical details. The tools only read information and cannot make changes to your account.

Company discovery

search_companies — Finds companies available in Dafinchi by name, stock symbol, keyword, industry, sector, or market value.
get_company_details — Looks up company profiles and available earnings periods.
get_available_industries_sectors — Shows the industries and sectors you can explore.
get_company_peers — Finds stock-market peers and companies with similar business profiles.

Earnings and yearly reviews

list_transcripts — Finds available earnings transcripts and their release dates.
semantic_search_transcript — Finds transcript passages relevant to a question or concept, even when the wording differs.
fulltext_search_transcript — Searches earnings transcripts for keywords or phrases.
get_earnings_analysis — Retrieves existing earnings analysis and summaries for companies and reporting periods.
list_yearly_reviews — Finds completed company yearly reviews published on Dafinchi.
get_yearly_review — Reads the content of published company yearly reviews.

Financials and market data

get_stock_price_data — Retrieves historical stock prices and trading volume.
get_fundamentals_snapshot — Shows recent financial statements, plus financial measures and ratios for the past 12 months.
get_financial_statements — Retrieves historical financial statements and product or business-segment revenue breakdowns.
get_consensus_snapshot — Shows analyst price targets, a summary of their expectations, and the latest rating.
get_insider_trading_snapshot — Summarizes recorded quarterly insider buying and selling activity.
get_revenue_breakdown — Shows how revenue is distributed across products or business segments.

Connection help

Problem	What to do
Cannot connect	Copy the full MCP server URL from Connection settings above and try again. If it still does not work, contact Dafinchi.
Access denied	Make sure you signed in to the right Dafinchi account and MCP access has been enabled for you. To request access or get help, contact Dafinchi.
Sign-in expired or failed	Reconnect Dafinchi from your AI application's settings, then sign in and approve access again in your browser.
The application asks for an API key	Choose OAuth or browser sign-in if available. Dafinchi MCP uses your Dafinchi sign-in, so you do not need to create an API key. If your application only accepts keys or local commands, check its help page for remote MCP support.
Usage limit reached	Wait before trying again and follow any retry instructions shown by your assistant. Your account's usage limits also apply when you use MCP.