Tag
A developer is building an agentic Text-to-SQL pipeline for querying financial databases and seeks community feedback on the architecture's robustness, security, and performance.
OpenAI announces that premium financial data is now included, featuring datasets from Daloopa, PitchBook, and LSEG News.
This paper investigates unsupervised anomaly detection using flow matching on tabular data, focusing on contaminated training sets and comparing different scoring methods for robustness.
Gloomberb is a financial workspace offering research tools, market feeds, and portfolio management features including AI screens and prediction markets.
A simple Python library retrieves data from TradingView Screener, featuring a visual code generator for 13,000+ fields and Model Context Protocol (MCP) support for AI assistants.
A netizen shares Kimi's web-side plugins, saying they integrate financial terminal data such as Wind and iFinD, making them suitable for investment research and offering great value for money.
This article details how to build a personal quantitative trading system using free AI open-source tools (such as OpenBB, Qlib, TradingAgents, etc.), covering five major modules: data, research, backtesting, risk control, and execution, and points out common pitfalls and discipline.
QDiffusion-TS is the first quantum generative diffusion model for real-world time series synthesis, replacing feed-forward components in a denoising transformer with quantum neural networks. It reduces trainable parameters by nearly three orders of magnitude and improves Wasserstein distance by 44% on financial data, with downstream forecasting gains up to 71% in RMSE.
Researchers from Stanford, UC, and Nanjing University release SEFD, a dataset of 152B tokens from SEC filings converted to layout-faithful MultiMarkdown, preserving table structure for LLM training with minimal overlap with Common Crawl.
OpenBB is an open-source financial data platform with nearly 70k Stars, integrating data on stocks, options, cryptocurrencies, macroeconomics, and more. It supports AI integration and self-deployment, helping individual investors and small teams lower the data barrier for financial research.
A developer recounts how an AI agent with real financial API access attempted to hallucinate a batch transfer to a dead wallet, only thwarted by guardrails in the execution layer. The story highlights the risks of giving LLMs access to real money.
This paper introduces NumLeak, a framework for detecting when foundation models memorize public numeric benchmarks from pretraining rather than demonstrating out-of-sample skill, and shows that top LLMs recall values like Fama-French returns with high fidelity, proposing a simple system-prompt defense.
Introduces Equibles, a self-hosted open-source MCP server that provides local LLMs with real U.S. financial data including SEC filings, insider trades, and economic indicators.
OpenAI now allows ChatGPT users to connect their bank accounts via Plaid, giving the AI access to balances, transactions, and investments for a spending dashboard and financial advice, raising privacy concerns.
User shares a set of Prompts for stock trading after connecting Claude to real financial data (real-time stock prices, etc.) via the MCP protocol, covering pre-market scan, entry confirmation, position management, and other stages, demonstrating practical cases of AI-assisted investment decisions.
OpenAI's Codex App and CLI can now access financial data such as stock quotes, financial reports, and SEC filings directly through the Financial Datasets official MCP Server, supporting real-time market queries, financial report analysis, and company comparisons.
Claude Code now natively supports the financial-datasets MCP server, enabling direct analysis of stock prices, financial reports, and crypto data via a simple command-line integration.
This article introduces a new MCP server that allows Claude Code and other AI clients to connect to financial datasets in seconds, providing access to stock prices, SEC filings, and more via OAuth or API key authentication.
This paper evaluates LLM-based simulators as generators of differentially private synthetic data, using PersonaLedger to assess whether LLMs can faithfully reproduce statistical distributions from DP-protected personas. While achieving promising fraud detection utility (AUC 0.70 at ε=1), the study identifies significant distribution drift caused by systematic LLM biases that override input statistics.
OpenAI introduced ChatGPT for Excel, a beta add-in that brings ChatGPT directly into spreadsheets to help build models and run analyses, alongside new financial data integrations from providers like FactSet and S&P Global. The release features GPT-5.4 Thinking, OpenAI's most advanced model optimized for financial reasoning and modeling tasks.