how to build an ai trading bot with deepseek

turn ai into a crypto trading powerhouse. in this guide, i'll show you step by step how i built an ai-powered trading bot that turned $100 into $35k—all using deepseek ai and dexscreener api.

this isn’t clickbait. the bot runs 24/7, analyzes the market, and trades automatically. by following these steps, you can build your own.

what is deepseek ai?

deepseek recently released r1, an open-source ai model with powerful capabilities similar to chatgpt and claude—but completely free.

i used it to build a trading bot that:
✅ connects to dexscreener api
✅ analyzes the market in real-time
✅ filters out scams & rug pulls
✅ executes trades automatically

performance results

in just 18 hours, the bot caught multiple high-momentum trades:

• csi300: 35x

• fartfrog: 22x

• januspro7b: 15x

• h5n9: 12x

while it wasn’t trading, the bot was learning, building a database, and refining its strategy to make fewer mistakes.

step 1: setting up deepseek ai

to create this bot, you’ll use deepseek chat—it’s free to use.

💠 go to chat.deepseek.com
💠 use this prompt:

"hello. please pretend you are a professional programmer with a harvard degree and 10 years of experience. create a bot to interact with dexscreener. we need to parse, save, and analyze each coin that rugged/pumped/become tier-1, listed on cex, etc to find the patterns."

step 2: adding filters & blacklists

the bot isn’t useful unless it can filter out bad coins. we need to add:

• filters to track good opportunities

• a blacklist to avoid scam coins

• a dev blacklist to avoid rug devs

💠 use this prompt:
"modify the code to use filters, blacklist of coins, and dev blacklist to blacklist rug devs. and add all of these settings into the config file."

step 3: avoiding rug pulls

to make sure the bot doesn’t buy into fake pumps, we need to check coins for fake volume before trading.

💠 use this prompt:
"avoid coins with fake volume, check if it's fake via your algorithm or use pocket universe api that works with dexscreener too."

step 4: extra safety checks

to make it even safer, we’ll add rugcheck.xyz and check for bundle purchases to avoid locked liquidity scams.

💠 use this prompt:
"check each token on http://rugcheck.xyz and interact only with contracts marked as 'good.' also, check if the supply of the coin wasn't bundled. if bundled, blacklist + dev blacklist."

step 5: automating trades

to execute trades, you’ll need to connect your bot to a trading system. the easiest way is to use bonkbot or trojan via telegram.

💠 use this prompt:
"to trade selected tokens, use bonkbot via telegram. also, add tg support to get notifications of buy/sell, and merge all code into one file. also, write how to launch this code."

if you want a ui for your bot, you can ask deepseek to generate one. it successfully created a ui for mine.

ready to launch?

your ai trading bot is now fully set up. it can analyze trends, filter out scams, and trade automatically.

click here for full source code

Python 3.12.8 ONLY.

this is just the beginning—ai trading is the future. time to start building.