Presentation of AlphaDesk - 30th May 2025
A unified interface for quant finance, inspired by FinViz, TypingMind, and Notion - Roadmap to December 2025
This document outlines the development of AlphaDesk, a self-hosted, data-centric research environment for quantitative finance. Inspired by platforms like TypingMind and Notion, AlphaDesk seeks to provide a customizable low-code workspace for agentic workflows and structured financial analysis. The long-term objective is to build a modular system that supports equity analysis, strategy development, and academic research through an intuitive project-based interface.
This log is focused strictly on documenting development progress. For equity-specific or strategy-driven research, readers are directed to the author's Medium page or the archive section of previous platform iterations.
Vision
AlphaDesk is conceived as a unified interface for managing quantitative research workflows. Key principles include:
Project-based structure: Each research unit is scoped as a standalone project (e.g., stock analysis, portfolio simulation).
Data persistence: All project data is stored locally in structured JSON format.
Agentic intelligence: AlphaDesk will support GPT-based agents for research, analysis, and visualization.
Customizability: Users can extend core functions through modular, low-code components.
The roadmap includes four core project types:
Phase 1: Stock Fundamentals and Technical Analysis
The first implementation focuses on individual equity studies. When a stock project is created:
End-of-day (EOD) historical data is fetched and visualized.
Fundamental data is normalized and displayed in structured panels.
News articles (not shown in the initial screenshot) are retrieved and rendered.
All retrieved data is automatically persisted to disk. This enables re-usability and the foundation for cross-sectional studies. A local JSON store acts as a growing database for downstream tasks.
Phase 2: Portfolio Construction and Risk Assessment
The next project type will focus on multi-asset analysis:
Multiple ticker EOD data is fetched.
Portfolios are simulated historically.
Monte Carlo simulations model future equity curve paths.
Quantitative metrics such as return distributions, beta regression, and cumulative performance will be introduced.
This extends functionality previously implemented in "Chat with Fundamentals," and will serve as a sandbox for exploring dynamic risk assessment models.
Phase 3: Strategy Design
This module will build upon prior work with QuantCoderFS, allowing:
Aggregation of multiple research articles.
Extraction and comparison of trading features.
Encoding of strategy logic into reusable modules.
The aim is to automate early-stage research and prototype validation.
Phase 4: Quantitative Research
This project category focuses on academic and theoretical studies:
Leveraging GPT-based agents to summarize literature.
Formalizing hypotheses from papers.
Generating annotated references and citations.
By integrating GPT-researcher capabilities into AlphaDesk, users will be able to conduct structured reviews, replicate findings, and build a foundation for original work.
Phase 5: Low-Code Interaction Agents
The fifth phase introduces a framework for defining and storing low-code agents that can manipulate and analyze data through custom workflows. These agents are intended to function similarly to those in OpenAI GPTs or TypingMind, enabling:
Personalized automation of data analysis tasks.
Integration of logic-based decision routines into projects.
A consistent interface for invoking reusable, user-defined workflows.
At this stage, implementation remains conceptual, with no baseline code yet available. However, this functionality is envisioned as central to AlphaDesk’s long-term goal of agentic, user-driven automation.
Release and Development Timeline
The first public release of AlphaDesk is tentatively scheduled for December, contingent on development progress. The platform's data pipeline is powered by EODHD, and access to their API is required. Interested users may register via the affiliated links :
EODHD - Fundamentals data feed
Development activities will pause during the summer and resume in September, in line with the guiding principle:
"Stop coding in May and go away."
Community feedback on features, workflows, and overall design is both encouraged and appreciated. The ultimate goal is to establish AlphaDesk as a robust, adaptable tool for both independent researchers and institutional quantitative analysts.