Open-source Python data platform that connects financial and other data sources and exposes them to Python, Excel, AI agents and REST APIs.
virattt/dexter
An autonomous AI agent for deep financial research that plans tasks, self-checks its work, and pulls live market data.
Open-source alternative to:Bloomberg Terminal
- Stars
- 27,761
- Forks
- 3,421
- Open issues
- 106
- Last commit
- last month
What is it?
Dexter is a terminal-based financial research agent written in TypeScript and run with Bun. Given a financial question, it breaks the request into structured research steps, selects tools to gather data, checks its own intermediate results, and iterates until it produces a data-backed answer. It is positioned as a Claude Code-style agent specialized for finance.
The agent has access to financial data such as income statements, balance sheets, and cash flow statements through the Financial Datasets API, and can run web search through Exa or Tavily. It supports several LLM providers, including OpenAI, Anthropic, Google, xAI, OpenRouter, and local Ollama models. Safety limits such as loop detection and step caps guard against runaway execution.
The project also includes an evaluation suite that scores the agent against a dataset of financial questions using LangSmith tracing and an LLM-as-judge approach, plus a scratchpad that logs every tool call to JSONL files for debugging. An optional WhatsApp gateway lets you send questions to the agent from your own chat. The project is explicitly for educational and informational use and is not intended for real investment decisions.
Key features
- 01Breaks complex financial questions into step-by-step research plans
- 02Pulls income statements, balance sheets, and cash flow statements
- 03Checks its own results and iterates until the answer is complete
- 04Supports multiple LLM providers, including local Ollama models
- 05Logs every tool call to scratchpad files for debugging
- 06Includes an evaluation suite with LangSmith tracing
Use cases
- Researching a public company's historical financial statements from a terminal
- Studying how an AI agent plans and uses tools for data-heavy analysis
- Benchmarking LLM agents on financial question datasets
- Chatting with a financial research agent through WhatsApp
Quick start
- 1
Install the Bun runtime (v1.0 or higher) from bun.com
- 2
Clone the repository
git clone https://github.com/virattt/dexter.git - 3
Install dependencies with Bun
bun install - 4
Run Dexter in interactive mode
bun start
Things to know
- Requires an OpenAI API key and a Financial Datasets API key, which may carry costs.
- The README says it is for educational and informational use only and is not meant for real trading or investment.
- Outputs may be incorrect or out of date, so verify results before relying on them.
Topics
FAQ
- What is dexter?
- An autonomous AI agent for deep financial research that plans tasks, self-checks its work, and pulls live market data. Dexter is a terminal-based financial research agent written in TypeScript and run with Bun. Given a financial question, it breaks the request into structured research steps, selects tools to gather data, checks its own intermediate results, and iterates until it produces a data-backed answer. It is positioned as a Claude Code-style agent specialized for finance.
- Is dexter free to use? What license does it have?
- dexter has no declared license. Its code is public, but without a license you have no explicit permission to reuse it; ask the author before relying on it.
- Is dexter still maintained?
- Updated within 6 months. The last commit to dexter was on Sep 23, 2026.
- What can I use dexter for?
- Researching a public company's historical financial statements from a terminal Studying how an AI agent plans and uses tools for data-heavy analysis Benchmarking LLM agents on financial question datasets Chatting with a financial research agent through WhatsApp
- How do I install dexter?
- Install the Bun runtime (v1.0 or higher) from bun.com Clone the repository: git clone https://github.com/virattt/dexter.git Install dependencies with Bun: bun install Run Dexter in interactive mode: bun start
- Where can I get dexter?
- The source code and installation instructions are on GitHub: https://github.com/virattt/dexter
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source finance terminal for quotes, charts, financials, filings, portfolios and alerts, run from your terminal or a desktop app.
Native C++20 desktop terminal for financial research, with Qt6 UI, embedded Python analytics, AI agents and 100+ market data connectors.
A Rust-backed Bloomberg client with Python and Node.js bindings for reference, historical, intraday, streaming and BQL data requests.
A terminal stock dashboard in Rust that pairs live quotes and charts with AI-scored news, SEC Form 4 insider filings and earnings reads.
A collection of plug-and-play agent skills that give LLM agents finance capabilities: news, stock data, sentiment, forecasting and reporting.