How I Built Three Passive Income Bots on My Phone in One Afternoon — No Coding Experience
Six months ago, if you’d told me I’d be running three automated investment bots that manage my portfolio while I sleep, I would have laughed. Not a polite chuckle — a proper, genuine laugh. Because here’s the thing: I don’t know how to code. I’ve never taken a computer science class. My finance knowledge came entirely from YouTube videos and the occasional Reddit rabbit hole.
And yet, here I am. Founder of PocketBots. Running automated systems that rebalance my portfolio, harvest dividends, and scan cryptocurrency exchanges for arbitrage opportunities. All built on a single afternoon, using nothing but my Android phone and a healthy dose of stubbornness.
This is the real story of how it happened — including the bits where I nearly gave up, the moments where I genuinely couldn’t believe it was working, and everything I learned along the way. I’m sharing this because I want you to know that if I can do this, you absolutely can too.
The Frustration That Started Everything
Let me paint the picture. It was a rainy Saturday in March. I was sitting in my flat, scrolling through my investment accounts, feeling that familiar sense of overwhelm. I had a bit in stocks, some crypto scattered across two exchanges, a small bond allocation I’d set up years ago and forgotten about. Nothing was organised. Nothing was optimised. I’d read about portfolio rebalancing and dividend investing strategies, but actually implementing them felt like a second job.
Every few months, I’d manually log into each account, calculate my current allocations, figure out what needed buying or selling, and then inevitably get distracted halfway through. My portfolio was a mess, and I knew I was leaving money on the table.
That Saturday, I decided I was done with manual management. I’d heard about trading bots and automation, but everything I’d seen required either expensive software subscriptions or serious programming knowledge. I had neither. What I did have was my phone, a stubborn refusal to accept that this was too complicated for me, and a vague awareness that AI assistants had become remarkably capable.
Meeting Claude: My Unexpected Coding Partner
I’d used ChatGPT a few times for random questions, but I’d recently heard about Claude AI and how it was particularly good at helping with technical tasks. So I opened it up on my phone and typed something like: “I want to build an automated bot that rebalances my investment portfolio. I have no coding experience. Can you help me do this?”
What happened next genuinely surprised me. Instead of giving me a generic overview or telling me to hire a developer, Claude started asking specific questions. What broker did I use? What assets did I want to include? How often did I want rebalancing to occur? It was like having a conversation with a patient tutor who happened to be an expert in everything.
Within twenty minutes, Claude had explained the basic architecture I’d need: a small server to run the code, a programming language called Python (which I’d heard of but never touched), and something called API connections that would let my bot communicate with my brokerage accounts.
It sounded intimidating. But Claude broke it down into steps so simple that even I could follow them. And here’s the key insight I want to share: you don’t need to understand everything to make it work. You just need to follow the process, one step at a time.
Setting Up the Server (From My Phone, in Termius)
The first practical step was getting a server — basically a small computer that runs 24/7 in the cloud, executing my bots even when my phone is off. Claude recommended DigitalOcean, which I’d never heard of but turned out to be perfect for beginners.
I signed up on their website, and within five minutes, I’d created what they call a “Droplet” — a tiny Ubuntu Linux server that costs just £5 per month. Less than two cups of coffee. This server would become the brain running all my automation.
Now, here’s where it got interesting. To actually work with this server, I needed a way to connect to it and type commands. On a laptop, you’d use a terminal application. But I was doing everything on my phone, so I downloaded an app called Termius from the Google Play Store.
Termius is essentially a terminal emulator for mobile — it lets you connect to remote servers and run commands using a mobile-friendly interface. I’ll be honest: the first time I connected to my DigitalOcean server through Termius and saw that blinking cursor waiting for input, I felt like a hacker in a film. I had no idea what I was doing, but it felt exciting.
I copied the connection details from DigitalOcean — the IP address, username, and password — pasted them into Termius, and suddenly I was looking at my own server. A blank canvas waiting for instructions.
Building Bot Number One: The Portfolio Rebalancer
With Claude guiding me through every command, I started building my first bot. The concept was simple: I wanted to maintain a portfolio allocation of 60% stocks, 25% cryptocurrency, and 15% bonds. Whenever the markets moved and my allocation drifted away from these targets, the bot would automatically buy or sell small amounts to bring everything back into balance.
Claude walked me through installing Python on my server — just a few commands I typed into Termius. Then it helped me create a Python script, explaining each section as we went. The script would:
- Connect to my brokerage API to check current holdings
- Calculate the current percentage allocation of each asset class
- Compare that to my target allocation
- Generate buy or sell orders for anything that had drifted more than 2% from target
- Execute those orders automatically
I won’t pretend I understood every line of code. I didn’t. But I understood the logic, and Claude patiently explained anything I asked about. When something didn’t work — which happened several times — I’d paste the error message back to Claude, and it would explain what went wrong and how to fix it.
The first time I ran the rebalancer and it actually executed a trade — selling a small amount of stock that had grown beyond its target allocation — I genuinely gasped. It worked. My phone had just built something that managed my investments.
Bot Number Two: The Dividend Harvester
Emboldened by my success, I moved on to the second bot. This one was focused on dividend investing — specifically, building a position in high-yield ETFs and tracking the income they generated.
After researching dividend ETFs (again, mostly through conversations with Claude), I settled on four funds: SCHD, VYM, JEPI, and HDV. Combined, these give roughly a 4.6% annual dividend yield, paid out quarterly. Not life-changing money on a small portfolio, but proper passive income that compounds over time.
The dividend harvesting bot was simpler than the rebalancer. Its job was to:
- Track dividend payment dates for each ETF
- Automatically reinvest dividends when they arrived
- Send me a weekly summary of projected annual income
- Alert me if any fund’s yield dropped significantly
Setting this up took about forty minutes. The trickiest part was configuring the API authentication with my broker — there were several steps involving something called OAuth that confused me completely. But Claude explained it like I was five, and eventually, I got it working.
Bot Number Three: The Crypto Arbitrage Scanner
The third bot was the most ambitious — and honestly, the most fun to build. Crypto arbitrage is the practice of exploiting price differences between exchanges. If Bitcoin is selling for £50,000 on Kraken but £50,150 on Coinbase, there’s theoretically a profit opportunity in buying on one and selling on the other.
Now, I want to be clear: arbitrage opportunities in major cryptocurrencies are rare and usually tiny. The big hedge funds with supercomputers snap them up in milliseconds. I wasn’t expecting to get rich from this bot. But I wanted to see if it could find opportunities and understand the mechanics.
Claude helped me build a script that:
- Connected to both Kraken and Coinbase APIs
- Checked prices for Bitcoin, Ethereum, and a few other major coins
- Calculated the price difference as a percentage
- Logged any difference greater than 0.3% as a potential opportunity
- Ran automatically every 5 minutes
Setting up API keys for both exchanges was straightforward — both Kraken and Coinbase have good documentation, and Claude helped me navigate it. The bot doesn’t automatically execute trades (I’m not that brave yet), but it logs opportunities and sends me notifications.
In the first month, it found seven instances where price differences exceeded my 0.3% threshold. Most were fleeting and would have been eaten up by transaction fees anyway. But two were substantial enough that, if I’d been faster, I could have made a small profit. It was proof of concept — and genuinely exciting to watch.
Automating Everything with Cron Jobs
Here’s where the magic really happened. Building the bots was one thing, but having them run automatically without my intervention was another. That’s where cron jobs came in.
A cron job is basically a scheduler for Linux systems. You tell it when to run something, and it just… does it. Forever. No further input required.
Claude taught me the syntax — which honestly looks like hieroglyphics at first — and helped me set up schedules for each bot:
- Portfolio Rebalancer: Runs every Sunday at 6pm
- Dividend Harvester: Runs every Monday morning at 8am
- Arbitrage Scanner: Runs every 5 minutes, 24/7
Once the cron jobs were configured, I could close Termius, put down my phone, and forget about it. My little £5/month server was working for me around the clock.
What Went Wrong (And How I Fixed It)
I’d be lying if I said this was completely smooth. Here are the main obstacles I hit:
API Rate Limits: Early on, my arbitrage bot was pinging the exchanges too frequently and got temporarily blocked. Claude helped me add delays between requests and implement proper error handling.
Timezone Confusion: My server was set to UTC, but I’m in the UK. My rebalancer was running at the wrong time. A simple timezone configuration fixed it, but it took me an hour of confusion first.
Typos: I can’t tell you how many times a single mistyped character broke everything. Python is unforgiving about syntax. But Claude was patient, and eventually, I learned to double-check my code more carefully.
Authentication Expiry: My broker’s API token expired after 30 days. The bot just stopped working one morning. Now I have a reminder to refresh it, and I’m working on automating that too.
The Results After Six Months
Today, all three bots are still running on that same £5/month server. My portfolio stays automatically balanced, my dividend income is tracked and reinvested, and I get occasional alerts about arbitrage opportunities.
Has it made me rich? No. But it’s saved me hours of manual work, removed the emotional decision-making from my investments, and taught me an incredible amount about programming, APIs, and financial systems.
More importantly, it’s shown me what’s possible. If I could build these three bots on a Saturday afternoon with no coding experience, what else could I build? What could you