How To Build a Polymarket Copy Trading Bot With Zero Coding (With OpenClaw) transcript

What if you didn't have to predict the market yourself? What if you just automatically copied every trade from the best trader on Polymarket? And like instantly, every time they bet, you bet, too. I'm building that right now. We're going [music] to do it live, and you're going to follow along. A few weeks ago, I made a video showing how I built a Polymarket trading bot using OpenClaw. That video hit 21,000 views, and the bot is still running today.

But, people kept asking, "What's next?" So, I kept thinking about this. I mean, my original bot uses an AI to predict 15-minute crypto markets, and it works. But, the thing is prediction is hard. So, what if instead of predicting, you just copied someone who's already winning? And that right here is copy trading in a nutshell. You find the best trader on Polymarket, someone with a 60% winning rate, say, a consistent P&L, profit and loss, who's active right now, and you just point your bot at their wallet.

Every trade they make, your bot mirrors it proportionally, automatically. They bet $200, you bet $20. They sell, you sell. You don't think, you just follow. And that's exactly what we're going to do today. I'm going to show you how to build that using OpenClaw from scratch with zero coding. You will need two things before we get to the fun part. First, you need your Polymarket credentials, and second, you need to give those to your OpenClaw.

Then, you just describe what you want, and let it build itself. So, let's go. Let's do that together. These are your private keys. They're like the password to your wallet. So, please make sure never to share these with anyone other than in OpenClaw. But, that being said, I do recommend that you create a new wallet for this bot, mostly because in case it goes haywire, in case it causes mayhem, you're not risking your life saving capitals.

And once you do that, open your Open Claw, whether you talk to your Open Claw on Telegram, Discord, whatever it is, open it and give it your credentials so that it can use them when building your bot out. Just type something like store my Polymarket credentials, wallet address, paste your address, private key, paste your key, use these whenever you're building anything Polymarket related. Open Claw will store this securely.

It's on your local machine. It's not being sent anywhere. And from this point forward, whenever you tell Open Claw to build something on Polymarket, it already has what it needs. You never have to paste your credentials again. Okay, the setup is done. Now onto the fun part. So, the two things that you will need are first, an Open Claw setup. That's the AI that's going to do all the work. And if you have your Open Claw setup, then the second thing you'll need is a Polymarket account with USDC in it.

That's your trading capital. You can load it up as comfortably as you feel. If you want to go for the minimum, go for the minimum. Whatever your risk tolerance is. So, let's get your credentials. In order to do that, you're going to want to go to your settings. You'll need to go to builder codes, and you'll need to press on create new. This will give you your API key, your secret, and your passphrase. You will want to include these in your Open Claw prompt when you're setting the bot up yourself.

And here is the message that I sent Open Claw to start building the entire bot. I told it, "Build a Polymarket copy trading bot from scratch. Use only the official Polymarket clock client NPM package. No random Git- GitHub repositories. The reason I'm telling it that is from experience, I've learned that a lot of the popular GitHub, the most starred GitHub repositories, are just drainers and hiding in plain sight. So, just if you end up using one of those, just ask your open clock to scan it for any malicious activity or potential.

Uh but for this bot that we're building, I want to build it completely from scratch. I don't want it to rely on any repositories for fear of being drained. But that being said, credentials are already in the previous conversation context. So, I've already given it the credentials. Step one, scrape the Polymarket leaderboard to find the top traders by ROI efficiency. This is how I want it to filter for the top traders that we're going to end up copying.

Step two, build a bot that pulls those wallets for new positions and mirrors trades proportionally. This is its logic. Once it sees a trader make a trade, it will mirror that. And step three, start in preview mode. I want it to start paper trading, so not using any real money until we make sure that its logic is actually set up, it's actually working, it's an act it can actually do the job that we need it to. Log what trades would be made, don't execute anything real until I say go live.

Audit every dependency before installing. If any package looks suspicious or isn't from a known publisher, reject it. Let's get started. And with that, here's what it told me. So, good, we already have that dependency installed. Let me check the existing whale configuration and hit the Polymarket API. Let me grab wallet addresses for the top ROI traders by by clicking their profiles. So, it snatched all of the Polymarket traders it wants to copy.

I have everything I need. Now, building the bot. This is I absolutely love OpenClaw. You can throw any riddle, puzzle its way, any challenge its way, and it'll figure it out. It'll start breaking things down in a way that it can solve one problem after the other. And, just like that, it fetched five traders that are profitable, that have done some good volume, that have great ROI along with their wallets, and it's now building the bot.

Just like that. A single prompt, and it's now building the bot. Let's check in on it in a couple of minutes and see its progress. Okay, and just like that, it has built us a bot. What it does, it pulls all five wallets every 30 seconds for new activity. It filters no crypto markets, only buys trades must be 5 minutes old, or less than 5 minutes old, price between 0.6 and 0.95. Why is it pulling for a price between that?

Whale conviction stronger than $30. Size trades proportionally to ROI weight. Tracks profit and loss, active positions, daily budget, and win loss record. So, very nice. The only thing I'm wondering is why Let's actually ask it. Why is it filtering for a price between 0.6 to 0.95? Let's see what it says. Two reasons. Uh went ahead and outlined the reasons. I don't really agree with those reasons, so I ended up telling it, "No, let's copy whatever the top traders are doing." I mean, that's the purpose of the copy trader.

And then, eventually, if we find that those are the specific trades that aren't profitable, then we can just apply that filter again. Uh but, yeah, it told me it makes sense. Uh if this, whatever this is, fit to would be fit. It's betting something at 0.3. No, that's a trader. If he's betting something at 0.3, who are we to second-guess him? Exactly my point. Done. Price filter removed. It'll now copy at any price as long as the trade is fresh, it's a buy, and the whale is putting in at least $30 or more.

But with that, let's ask it to create a dashboard for us to be able to track its activity. So, let me go ahead and type that out. So, this is what I wrote it. Can you build me a dashboard so I can track the copy trading bot's activity that we just deployed? I'd like to be able to see everything we're doing from tracking the trades, the logs, the wins and losses, the P&L, and anything you think might be helpful. So, this is what you would copy and what you would write it yourself.

Now, I already have a mission control. So, in case you do as well, then you can ask it to include it inside your mission control. Ask it to include include it inside the trading tab of Max HQ. Maybe create a sub tab called copy. And boom, just like that, it's created a view for me to be able to watch the bot as it works. Here it is, the copy mode. Today's spend with the total P&L, our win rate, active positions, daily budget, and our tracked traders.

There's no trade history yet. Bot logs are empty so far. It hasn't made a trade, that makes sense. Says copy trader is offline. So, let's go check on whether it's online or not. Okay, and with that, it is officially now live in preview mode. So, no real trades are happening. Let's check its bot logs. There you go. So, it's checking every 30 seconds for new trades. Polymarket copy trader starting and it's looking at the different trades that are being made and checking for a suitable trade every 30 seconds.

Okay, so here's what I've done in the last maybe 10 minutes. I've asked it to add even more traders to our copy trading bot, the ones that it's tracking, so that we get more activity. This is me being impatient. I want more activity. And since we've done that, it literally just placed a trade right here. Uh Call of Duty Vancouver Surge versus Cloud9 New York. Who's the winner? It bet on Cloud9 at 29 cents. It copied this trader.

And we can see from its logs it's running every 30 seconds to check on new trades that it can make. But that being said, it just made its first bet. Now what we have to do is wait until the bet I need to fix this. It's showing it's lost, but the bet hasn't resolved yet. So let me check in on that right now. Oh, so that was actually a a trade that it made and it played and it resolved at a loss. And I ended up asking my open claw what happened with that specific trade.

Let me show you. And it told me that trade happened while it was fixing the 5-minute staleness filter. Why did it copy a resolved trade? Was the trade more than 5 minutes old from the trader we copied? And it basically went and troubleshooted why that trade happened and went ahead and fixed it. That being said, all I can do right now is just sit and wait. I need to be a little more patient. And I'll come back once we have some results we can actually look at.

So it's been 20 minutes exactly since we've went live and I've just asked it, how can we get more trades by copying more traders?" And it gave me a few options. So, yes, option number one is adding more traders. Option number two is trade is tracking higher frequency traders. And option number three is lowering the minimum conviction threshold from $5 to as small as two to $3. Dropping to $1 catches more of them. Uh it gave me a a fourth option, but I basically like the first three options, so I told it, "Yes, let's do 1 + 2 + 3." Let me scrape all pages at once.

Now, let me get pages two and three, then the weekly tab. And it's going ahead and doing that as we speak. But, this right here is exactly what I want you to take away from this. This is how easy it is for you to modify, pivot, work, build, shift, and work with your open claw agent. You tell it to your pain point. You highlight to it the problem and help it let you figure out a solution. I don't know any of this. I don't know how any of this works.

I don't know how copy trading works. I don't know how creating a polymarket trading bot works. I don't know how any of those APIs function. I don't know how to code. But, what I do know is how to talk to my open claw agent. And that is simple. That is easy. That's all you need to do. And this is the beauty of this. I mean, you can edit every single functionality that you have with your own agent. You can tell it, "Oh, add a button here for me to be able to add new wallets." Or maybe add a button where if I press it, you on your own go and fetch brand new fresh wallets that match my criteria.

There's There are so many things that you can do with this technology now. The sky is the limit. And with that, I'll see you in a few hours. >> Several hours later. >> All right, and just like that, it's been almost God, maybe 6 7 hours since we've last filmed. And in that time, we've managed to make $23. $23 with a 67% win rate and two active trades right now. We have Celtics versus Hawks and Nuggets versus Thunder, but with that, it notified me that our $20 budget was spent.

Time just went by quickly. And yeah, I mean, 67% win rate with $23 made out of an initial $20 in capital. I can't ask for more. So, what I'm going to do right now is just ask it to push it live. And just like that, we would have created our copy trader bot. We can see now that the $20 is exhausted. It just went offline on its own until the time resets and a new day starts. But, I'm about to tell it, "Hey, wake up and make us some more real money right now." And that right here is what AI agents actually are.

They're not chatbots. They're not tools you use. They're workers that you deploy. The skill now isn't coding. It's knowing what to build and being able to describe it clearly. And being able to have that back and forth, that constant back and forth with your AI agent. And that skill is a skill anyone can have. I'll post an update in a few weeks with the actual results, how much it made, who it copied, what went wrong, what I ended up learning.

I'm on the Claude Max subscription, which cost me about $100 a month, but you can use any AI model you want. Although, be warned, different AI models will show different results based on well, how smart they are. I do have one question before I let you go. What would you build if you had an AI that could just do anything? Drop it below in the comments. I'm curious. But that being said, subscribe if you want to see the results when this bot's been running for a month.

See you in the next video.