Co-Invest - Trade 500+ markets directly from ChatGPT & Claude
by•
Co-Invest by Liquid lets you research and place real trades inside ChatGPT or Claude. Access 500+ markets across crypto, equities, FX, commodities, and prediction markets—24/7. Fund in-chat, use on-chain signals (wallet activity, positioning, funding, order flow), and execute in plain language with stops/targets. Your AI also handles support with full context.
Replies
trading real money from a chat prompt is one of those things that sounds amazing until you think about what happens when you word something wrong. the confirmation step before execution matters a lot here
Co-Invest
@tina_chhabra The confirmation step is fully manual.
You said finance at its root is about asymmetric information where one side knows something the other doesn't what specific feature or piece of data inside Co-Invest today gives a retail user an informational advantage that even many small professional funds still lack access to?
Co-Invest
@carter_son Yup -- to name a few
24/7 access to prices (e.g. Crude Oil)
Proprietary on-chain positioning data
Access to the best analyst in the world -- Claude Opus 4.8
Really impressive scope here curious about the security model. When a user funds in chat and places a live trade, where are the keys or credentials held? Is execution going through a custodial wallet, a connected non-custodial wallet, or something else entirely? That's the piece I'd want to understand before putting real capital in.
Co-Invest
@daniel_juan2 Execution goes through a non-custodial wallet created on Liquid, a platform we also run that has 50,000+ users and $4 billion in volume. Importantly, even if we go bankrupt, your funds are safe.
Love the on chain signal layer. Quick technical question: are the wallet activity and order flow signals pulled in real time or is there a lag? For fast moving markets like perps, even a 30-second delay on funding rate data can meaningfully change the trade case. Would love to know the data freshness story.
Co-Invest
@new_user___10520260379921a76fc2d64 All data is as fresh as possible.
Congrats on the launch. Trying to understand the business model are you taking a spread on trades, charging a flat subscription, or is it fee per execution? With 500+ markets across so many asset classes, I imagine the revenue structure varies a lot by product type. Would be great to have more transparency on that upfront.
Co-Invest
@diego_joaquin1 We earn fees on trades.
The plain language execution is compelling. I'm wondering how the AI handles ambiguous instructions for example, if I say 'buy some ETH when it dips' does it ask for clarification, set a conditional order, or wait for explicit confirmation before doing anything? Understanding where the guardrails are would help build confidence in using it for larger positions.
Co-Invest
@ding_hao It is for explicit confirmation
Curious about the prediction markets integration specifically are you routing through an existing platform like Polymarket or Manifold, or have you built your own liquidity layer? Prediction market liquidity can be thin on non headline events, so I'd love to understand how you're handling slippage and order size limits there.
Co-Invest
@elena_fischer1 We use Polymarket for our prediction markets.
The AI handles support with full context feature is one of the most interesting parts of this product to me. Does that mean the support agent has read access to my full trade history and positions or just the current session? And is that handled by the same model or a separate system? The context boundary matters a lot for how much I'd trust it.
Congrats on shipping this. I'm trying to think through the liquidity side for equities and FX specifically, are you acting as the broker dealer directly, or routing through a prime broker or third party liquidity provider? That affects execution quality, margin requirements, and ultimately how competitive your pricing can be at scale.
Strong concept. Security question: if my conversation with Claude or ChatGPT contains trade instructions and account details what's the data retention policy? Is the trade context stored on Anthropic or OpenAI infrastructure on your servers or is it ephemeral per session? For institutional users especially that distinction matters quite a bit.