Autonomous AI agent that performs complex financial research by planning tasks, executing them with real market data, and self-validating results - like having a junior analyst that never sleeps

Ever wished you could ask “Should I invest in NVDA vs AMD?” and get back a comprehensive analysis with income statements, competitive positioning, and market trends? That’s exactly what Dexter does. Instead of manually gathering financial data from multiple sources and crunching numbers, you ask a question and watch this AI agent break it down into research tasks, execute them autonomously, and deliver data-backed insights.

What sets Dexter apart is its self-reflection capability - it doesn’t just run queries and hope for the best. The agent validates its own work, catches inconsistencies, and iterates until it’s confident in the results. It integrates with real financial APIs for live market data, has built-in safety features to prevent runaway execution, and can handle complex multi-step research that would take humans hours. The TypeScript codebase is clean and the setup is surprisingly straightforward with Bun.

With 14K+ stars, this is clearly resonating with developers who need financial analysis capabilities. Whether you’re building investment tools, conducting due diligence, or just tired of manually researching stocks, Dexter offers a glimpse into how AI agents can handle domain-specific research workflows that actually matter.

⭐ Stars: 14295
💻 Language: TypeScript
🔗 Repository: virattt/dexter