FinceptTerminal: a Bloomberg terminal built around 37 AI agents
A one-person finance app is trending on GitHub harder than anything else today. 1,169 stars in a single day. Total over 6,300.
FinceptTerminal is a native C++20 desktop app that tries to be a Bloomberg for the AI era. The trick is 37 built-in agents β Buffett, Graham, Munger, Lynch personas plus economic and geopolitical analysts β that you can chain through a visual node editor. Each one plugs into OpenAI, Anthropic, Gemini, Groq, DeepSeek or local Ollama, so you pick your brain per task and your wallet per run.
Underneath the personas there's actually serious machinery. DCF modeling, portfolio optimization, real-time trading across 16 brokers, and more than 100 data connectors from Yahoo Finance and FRED down to obscure government APIs. The MCP integration turns the whole node editor into a tool pipeline agents can drive. The v4.0.1 release on April 15 is what set the stars on fire.
The interesting part isn't the specific agents. It's the bet that retail-grade financial research gets repackaged as "talk to Buffett about this stock" rather than "read 80 pages of 10-K." That's a UI shift as much as an AI one.
Repo: https://github.com/Fincept-Corporation/FinceptTerminal
← Back to all articles
FinceptTerminal is a native C++20 desktop app that tries to be a Bloomberg for the AI era. The trick is 37 built-in agents β Buffett, Graham, Munger, Lynch personas plus economic and geopolitical analysts β that you can chain through a visual node editor. Each one plugs into OpenAI, Anthropic, Gemini, Groq, DeepSeek or local Ollama, so you pick your brain per task and your wallet per run.
Underneath the personas there's actually serious machinery. DCF modeling, portfolio optimization, real-time trading across 16 brokers, and more than 100 data connectors from Yahoo Finance and FRED down to obscure government APIs. The MCP integration turns the whole node editor into a tool pipeline agents can drive. The v4.0.1 release on April 15 is what set the stars on fire.
The interesting part isn't the specific agents. It's the bet that retail-grade financial research gets repackaged as "talk to Buffett about this stock" rather than "read 80 pages of 10-K." That's a UI shift as much as an AI one.
Repo: https://github.com/Fincept-Corporation/FinceptTerminal
Comments