13F.chat
13F holdings chat and AI portfolio research
About 13F.chat
13F.chat is a research platform built on Form 13F, the quarterly report in which U.S. institutional investment managers with more than $100 million in reportable U.S. equities disclose their long positions to the SEC. The filings are public, but reading them is slow work: each one is a list of CUSIPs, share counts and market values, and the interesting part, what a manager actually bought, trimmed or exited compared with the previous quarter, only appears once you line two filings up and separate share-count changes from price moves. 13F.chat does that processing for you and puts an AI assistant on top, so you can ask questions such as "What did Warren Buffett buy last quarter?" or "Which stocks are top managers buying together right now?" and get an answer grounded in the disclosed holdings.
The site is operated by LING QU AI LIMITED, a company registered in England and Wales, and positions itself as research for individual investors rather than advice. It is unusually direct about what 13F data cannot show: the filings are a snapshot published up to 45 days after quarter-end, they omit short positions, cash, most fixed income and non-U.S. listings, and they never reveal a manager's real purchase price. Its methodology page explains how simulations avoid look-ahead bias, and the about page warns that AI-generated answers can misread filings and should be checked against EDGAR. The product is actively developed, with a public version log showing several releases on 2026-10-24 alone, and research guides published in five languages. Pricing is freemium: a free plan with a small daily allowance of AI chats and fund analyses, and paid Plus and Pro plans that add monthly credit pools.
You can start without an account. Public research pages, such as the Berkshire Hathaway page, show the latest processed 13F snapshot for a manager: total reported value, number of positions, top-ten concentration, sector allocation and the quarter's new positions and exits, followed by a curated AI-written question and answer explaining what changed. The 13F Holdings directory lists funds and filers with return, alpha, Sharpe ratio, maximum drawdown, concentration and style labels, and can be sorted and filtered.
To go further you sign in and open the 13F Assistant or a workspace. The assistant is a chat interface with Enhanced, Basic and Research modes and a model selector; each chat costs 2 credits and a Fund Analysis run costs 5. You ask about a manager, a stock or a group of funds, and the assistant answers from the processed filing data.
Workspaces come in two forms. A Single-13F workspace focuses on one manager and offers tabs for Overview, Portfolio Radar, Insights, Factors, Portfolio, Cost Basis, Filing Pulse and a Strategy Lab. A Multi-13F workspace combines several managers, for example a Warren Buffett lineage pack of four portfolios, and highlights where they agree and disagree. From there you can paste your own holdings into Portfolio Insights or the optimizer to compare them with institutional positioning.
- •
13F Assistant - A chat interface for questions about managers, holdings and quarterly changes. Suggested prompts include what a manager bought last quarter, which names several managers are buying together, and whether a specific add is still worth following. Answers on public pages separate share-count changes from price drift and state the filing quarter and price date they rely on.
- •
Holdings Directory and Discovery Radar - The directory covers 92 13F filers and 114 funds, each with performance and risk metrics computed on a simulated post-disclosure basis. The 13F Discovery Radar aggregates the latest filings across 132 managers, 25 industries and 500 stocks, with a capital rotation timeline against SPY, a sector rotation heatmap, the largest industry rotations and the largest manager books.
- •
Single and Multi-13F Workspaces - Per-manager workspaces cover concentration, turnover, sector allocation versus a benchmark, factor exposure and an estimated cost basis. Multi-manager workspaces analyze several filers jointly and count consensus and disagreement positions across them.
- •
Backtests With a Disclosure-Day Rule - Simulations assume a reader acts at the next U.S. market open after a filing appears on EDGAR, rather than at quarter-end prices that nobody could have traded on. Results are benchmarked against SPY and reported as return, alpha, beta, Sharpe, Sortino, maximum drawdown, turnover and concentration.
- •
Portfolio Insights and Optimization - You can paste a portfolio to see factor tilt, sector concentration, single-name risk and overlap with tracked hedge funds, then run Mean-Variance, Black-Litterman or Risk Parity optimization. An AI strategy optimization stage tests challenger changes out of sample and only accepts those that clear a stated bar; the homepage reports 5 of 7 default workspaces accepted and 49 of 55 challengers rejected.
- •
Research Library - Reports and articles cover featured managers, ETFs such as POMG, JQTY and SEQ, comparisons like XLKM versus XLK, and scenario pages including portfolio-at-risk and tariff exposure. All static research reports are included on every plan.
13F.chat is aimed at individual investors who follow well-known managers and want to understand what their filings actually say without reading raw EDGAR documents. Someone who reads that Berkshire added to a position can check how large the change was, whether the weight moved because of buying or because of price, and how concentrated the portfolio now is, then ask follow-up questions in plain language.
It also suits more quantitative hobbyists who want to test whether following a manager's disclosures would have worked on a realistic timeline, compare several managers in one workspace, or run portfolio optimization anchored on institutional holdings. Finance students and writers covering fund managers can use the public research pages and methodology notes as a structured starting point, since each dataset carries its reporting period and update date.
It is not a source of real-time trading signals or personal advice. 13F data is delayed and incomplete by design, the site labels everything as research and education, and anyone using it should verify important figures against the original filings.
