Liquid launched Co-Invest on 26 May 2026, bringing account funding, market analysis and live trade execution into a single conversation within ChatGPT and Claude. The company describes it as the first product to combine all three without sending the user to a separate website or application.

The launch opens an important conversation for wealth managers, regulators and private investors. When an AI assistant can research an asset, size a position and prepare an order, the chat window becomes a natural point of contact between a client and a portfolio. Liquid’s approach places powerful analysis close to the decision while keeping the decision itself with the individual.

Trading inside the chat

Co-Invest gives users access to more than 500 markets spanning cryptocurrency, equities, foreign exchange, commodities, prediction markets and pre-IPO secondaries. Accounts can be funded by card, on-chain transfer or an external wallet, and users can set stop-loss and take-profit levels without leaving the conversation. The orders match what the Liquid mobile app offers, issued through chat.

Setup takes around a minute. Within Claude, the service is added as a custom connector, while ChatGPT users connect through the app directory and approve access to their Liquid account. There is no charge to install or use the product, no monthly plan and no surcharge for the AI layer, with users paying only the platform’s standard trading fees.

The workflow also scales beyond individual orders. An assistant can research a theme, suggest a basket of five trades and present them together, with a single tap from the user placing the full set. The launch followed the arrival of personal finance tools within ChatGPT earlier the same month, which allow users to connect bank accounts and portfolio data.

Professional insight for all

The thinking behind the product draws on Wang’s own background. Before founding Liquid, he led AI for Macro at Two Sigma, working where machine learning and financial markets overlap. His view is that professional investors have long benefited from better information, faster processing and more precise execution.

Co-Invest aims to make that toolkit widely available. The assistant can draw on positioning, funding, liquidation map and on-chain flow data, the type of information professional desks typically pay for. Wang has argued that AI does not replace judgement but makes it less expensive to be well informed, with execution as the natural next layer.

“Simplifying execution in the same interface was an obvious next step.”

Franklyn Wang Founder and CEO Liquid

Because the app runs inside ChatGPT and Claude, its research capability is set to improve as the underlying models advance. As Wang put it, “every time ChatGPT or Claude gets smarter, Co-Invest gets smarter as well.” Each new generation of AI model therefore strengthens the product it supports.

A clear regulatory line

The defining design choice concerns authority over each trade. The AI can analyse markets, recommend positions and prepare transactions, while a confirmation screen gives the user the final say on every order. Co-Invest does not execute trades on its own initiative.

That structure sets the product apart from an automated trading bot, which runs a strategy independently and places orders without further input. The assistant informs and prepares, and the client authorises. This clear division of roles gives investors the benefit of sophisticated analysis alongside full control of their capital.

Liquid also operates on a non-custodial basis, meaning users keep control of their own funds, and it routes orders to external venues including Hyperliquid. Wang has pointed to recent guidance from the Securities and Exchange Commission and the Commodity Futures Trading Commission, including a no-action letter issued by the latter to another platform, as supporting the company’s position that it does not need to register as a broker or exchange. The service is available in all 50 US states and most other jurisdictions, with sanctioned regions excluded.

Safeguards that build trust

Accuracy has been built into the architecture. The assistant handles the written response, while market inputs come from Liquid’s own data layer, including positioning data from Hyperliquid and headlines from authoritative sources. Combined with the confirmation step, this gives users reliable information and a clear moment to review each trade.

The conversational format also offers a fresh approach to customer protection. Financial apps commonly contend with impersonators posing as support staff, a particular consideration where crypto transactions cannot be reversed. With Co-Invest, support is provided within the conversation itself, and Wang has suggested the assistant would probably ask a number of questions before a very large transfer, citing a case in which an employee’s grandmother lost about a million dollars.

Scale and backing

Liquid enters this market with measurable traction. The platform has processed more than $3 billion in trading volume from around 40,000 users since launching in August 2025. April saw the company close an $18 million Series Seed extension co-led by Left Lane Capital and Neo, taking total funding to about $25.6 million, with Paradigm having led the earlier seed round.

Co-Invest also opens a valuable new route to customers. The existing user base arrived largely through crypto-native communities, whereas the new app reaches people who already use AI assistants to think through their investments. Wang expects the next generation of retail investors to begin their journey with AI, for example, going to their AI assistant with a question.

The company is currently prioritising the number of distinct users, with monetisation to follow. As AI assistants take on a larger role in how wealth is researched and managed, Liquid has set out an attractive model for that future.