Automating fishing for fun and no profit

I decided to automate a fishing game because I wanted to automate something, it was fun

Last updated: June 22, 2025

TL;DR

I decided to write a script that played a fishing game for me because I can, now I'm probably one of the few persons in the world to have completed it in its entirety without the use of mods

Introduction

A couple of weeks ago, I was browsing YouTube, as one does, and stumbled upon a video about a guy who decided to automate a few different Minecraft farms and essentially destroyed that server's economy. I really liked the idea. I'd seen it done a couple of times before, but this time I thought it would be a fun project. This got me thinking about trying automation myself, so I went to my Steam library and started scrolling through my games to see if I could find anything that would be fun to automate.

My goal was to find a game that was easy to automate, mainly through replicating mouse and keyboard actions, since most games don't let you pull out their data directly. Plus, I wanted to reach some kind of goal or achievement by automating it, rather than just automating for the sake of it.

That's when I stumbled upon Webfishing. Webfishing is, as you might have guessed, a game about fishing, it describes itself as:

WEBFISHING is a multiplayer, chatroom-based fishing game created by lamedeveloper. The game offers a cozy environment where players can catch nearly 100 different species of fish, meet new people, customize their character to their liking, play instruments, and generally be a silly little critter.

Webfishing is a sandbox game where 1-12 players can fish, metal detect, and play guitar on small online servers. Players earn cash by selling fish and completing quests, which they use to buy avatar customizations and equipment upgrades. The fishing minigame involves reeling in catches and occasionally clicking when prompted. In the game, you can reel in 79 different things. This includes 32 freshwater fish, 32 saltwater fish, 4 fish that only appear when it's raining, 7 undesirable junk items, and 2 other unique fish. Each time a player catches something it can be in one of 6 different qualities:

Normal
Shining
Glistening
Opulent
Radiant
Alpha

Each of these has a different probability of dropping depending on the type of bait used. Different baits significantly modify your chances of catching higher quality fish, with premium baits offering much better odds for rare Alpha quality catches. My objective with the game was to complete the fishing journal, for which I would have to catch all 79 different catches in each of the rarities, so a total of 474 unique items, each of them having a random chance that was determined by the rarity and quality of the catch.

The Plan

After playing for a bit I found that the game would be the perfect target for automation. The actions were simple enough, the objective was clear. From the Steam Achievements statistics I could see that 0.5% of players had actually completed the journal. Considering that there are mods that allow you to catch any fish you want, the number of players that had actually completed the journal "legitimately" was probably extremely low.

So I chose to write a script that would play the game for me, sounded simple enough. The main gameplay loop of the game is the following:

Select the type of bait
Go to a fishing spot you'd like to fish at
Cast your line
Wait for the fish to bite
Once the fish bites, you are prompted to a minigame where you have to cycle between holding and mashing the click button to reel in the fish in a certain amount of time.
If you are successful, you catch the fish and get the item, if you fail, you can try again.
Bait is consumed when you cast your line, so every time you run out of bait you have to go back to the shop, sell the fish you have, and buy more bait.
Selecting different types of bait will modify your odds of catching different quality fish.
The different fishing spots have different pools of fish, so you have to go to the ones that have the fish you want.

Here's a minimalistic recreation of the fishing loop

Fishing Minigame

With just a few additional details to work out, I was ready to start coding.

Automating webfishing

My first step was to tackle the fishing loop itself, since that would be the core of the script and the easiest to iterate over.

So first thing I did was to create a script that would cast a new line at max distance, then when the prompt for having a successful catch came up, it would hold the left mouse button down, reeling in the fish.

I had to figure out when to rapidly click the mouse button versus when to hold it down. At first I thought maybe I could get away with constantly mashing since each click would add a bit to the reeling progress, but I quickly realized that the game would not register as many clicks as I was clicking, meaning that it would only slightly move and eventually fail to reel in the fish.

My next approach was to hold until I saw the prompt for mashing. At that point click for a couple of cycles, then release the mouse button and wait for the next prompt. This worked okay for a bit but it had 2 issues. For higher quality fish it sometimes needed more clicks than the ones I programmed it to mash for, meaning it would get stuck holding when it had to mash and lose enough time to fail the catch. For lower quality fish it would mash way too much for no reason, eventually also leading to not reeling enough to catch the fish. For a human player this is quite easy since they can see the number of clicks required to mash, and count while they do so, but if I wanted to do it for a script I would need to figure out a way of detecting the number, reading it with computer vision, and then clicking for the correct number of times.

Catching a fish in the game — How it looks like when you catch a fish

Here I tried a bunch of different solutions, from using Tesseract to read the number, to trying opencv. The problem was I'd have to run this check after finishing every mash interval which could be many during a single catch. Parsing the image to get the number would take too much time, causing the reeling to fail. Speed was critical because the fishing mechanic has a tight time window - any delay in processing would result in a failed catch.

So in the end after a couple of hours of trying different solutions I found that there was a much easier way to do it. The game has a setting that allows you to switch the mashing mechanism to also work while holding the button down, so I just had to click once and then hold the button down.

This worked great, I could just click once and then hold the button down and the game would do the rest.

After the fish has been caught next thing is to skip all the text prompts for having caught the fish, which are either 1 or 2 depending on if you were lucky enough to also catch a treasure, and then the script could loop back and start fishing once again.

With that out of the way, the next challenge was changing the bait after it had run out. The game gives you a series of different baits you can have stocked at the same time, so my initial solution was to have it switch between them once it ran out. This worked okay and the script could keep running for a couple of hours without intervention. I almost thought of leaving it like that, but seeing how little of a dent it was making in the journal completion, it was clear I had to improve it and make it fully autonomous.

The first thing I did was to add a shop class that would handle the buying and selling of baits, and a fishing spot class that would handle the selection of the fishing spot, and the casting of the line.

Automating the whole game

Navigation and Store Integration

With the fishing loop done, the next logical step was to automate the process of selling fish and buying bait. The only problem with that is that I would also need to automate the process of getting to the store (a physically different location than the fishing spots) and then getting back to the same spot to guarantee the script would be able to keep fishing. If the position was off, over the iterations it would eventually be in a completely different place and the script would fail.

Luckily the game has a "Return to Spawn" button that could teleport me to a set location. With this I only had to write logic to move from the spawn to the store, and from the spawn to a fishing spot without having to worry about slight deviations as they would be fixed by the teleport. In the process of trial and error, I found that the camera did not reset to the default position when teleported, so I had to write logic to move the camera to a known position so the inputs which are camera relative would be correct.

This process took some time but eventually I implemented the logic to move to the store and to three different fishing spots which should be able to cover all fishing pools

Below are the routes to the freshwater, saltwater and void fishing spots respectively

With this done I felt confident that the script would soon be running continuously, so I went to sleep.

Optimization and Fishing Buddies

To my surprise, when I woke up the next day, the script was still running which was good, but when I checked the progress I found that progress on the journal was much slower than I expected. So I started looking for possible optimizations that could help me finish the journal faster.

I remembered the game had a "Fishing Buddy" feature which are basically 2 buckets one of freshwater and one of saltwater, and you could set them to automatically catch fish for you.

Fishing buddies — Setting up the fishing buddies in the void

They sounded interesting and after a bit of fiddling I managed to set up locations for them on all the fishing spots. After completing every second reeling attempt the script would now check on the fishing buddies to see if they had caught anything, this meant that on the time that before I got 2 fish, now I was getting 4.

More idling attempts, more fish, and the script was making good progress on the journal. I had already reached max level and bought all upgrades in about 50 or so hours of idling. At this point I wanted to make changes to the script so I could target the elements I needed to catch. As I said before each of the 79 items has a random chance of dropping in one of the 6 qualities, each type of bait modifies the odds of getting these qualities, up until now I was switching between baits as they ran out, but now I specifically wanted to target Alpha quality fish. So I added the option for picking which bait to use, and instead of switching to the next one it would instead initiate the buy/sell process when the current bait runs out.

Targeting Specific Fish and CLI Improvements

Around this time I also chose to polish a bit the CLI, with the different options for bait, fishing locations so that I could switch between modes once I was done looking for specific fish.

CLI with different options for bait, fishing locations and more

The stats

Before diving into the final results, here's what the script accomplished (it took a lot longer than I expected):

474 unique items collected (79 creatures × 6 rarities)
189 hours of fishing
15514 fish caught
3+ million in profit
4+ million exp more than required for max level

Victory

With that done, completing the journal became inevitable. I left the script running at night during a couple of days and eventually I woke up to a completed journal.

This was quite the different project from what I usually tackle but it was really fun, tedious at times to get the right set of inputs but overall it was a great experience, I'm glad to have done it, and there is some sense of accomplishment in knowing I 100% completed the game without the use of mods, even though I wouldn't call it completely legitimate because I used a script to do it. I still took the time to create and refine the script.

What I learned

This was a fun project that scratched the automation itch I've had for a while, not much learning in terms of new technologies, but it's fun to see some programming cornerstones being as important as ever.

Incremental improvement - Starting with basic fishing and gradually adding features was more effective than trying to automate everything at once
Defining self-contained flows - Breaking down the problem into smaller, manageable parts and testing each part in isolation, then orchestrating them to achieve the overall goal.

Automation continues to be one of my favorite ways to solve interesting problems, and this project reminded me why I love building things that work while I sleep!