PyOT

[~X~]

PyOT — A Modern Open Tibia Server, Reimagined in Python​

GitHub: GitHub - Mystikil/PyOT (https://github.com/Mystikil/PyOT)

---

What Is PyOT?​

A while back Stain started a Python distro called PyOT. It is a full-featured Open Tibia server completely written in Python 3 — not TFS, not C++, not glued with Lua. It’s a clean slate. Then it was latter mentioned by Amirosolo. SO credits go to them for the start of PyOT. I simply plugged in my custom Distro with theirs so far.
It was born from a simple idea:

“What if we could build a Tibia-like MMO server using clean, modern code that anyone could read, understand, and expand?”

That’s what PyOT is all about — a moddable, readable, Python-powered server core.
No compiler. No memory leaks. No guessing.
Just files, folders, and code you can actually understand.

---

Philosophy​

Traditional OT Servers	PyOT Approach
C++ base with Lua scripts	100% Python stack
Difficult to debug or modify	Readable, live-editable code
Statically compiled	Hot-reloadable scripts
Scattered logic	Modular folder-based systems
Slower development	Rapid iteration, even live

---

Core Features​

Gameplay Systems​

System	Status	Notes
Character Login / Accounts	Working	MySQL-backed
Movement & Pathfinding	Working	Includes monster follow behavior
Spells System	Modular	Python-defined, easy to extend
Vocations	Configurable	JSON or Python definitions
Monsters & AI	Basic	Advanced AI in progress
NPCs	Functional	Keyword & trade logic
Items System	Supported	Wearables, effects, usable logic
Combat / Stats	Functional	Cooldowns, damage types, scaling

---

Data Handling​

Feature	Description
JSON/Python Configs	Monsters, vocations, spells, etc.
Unified Loader	Drop data into folders, it's read at runtime
Hot Reloading	Modify systems without restart
Stat Scaling	Per rarity, level, or config values
Dynamic Item Definitions	Add new effects/bonuses easily

---

Maps & World​

Map Feature	Status
OTBM Support	Fully functional
Instanced Maps	In progress
Region Chunks	Base implemented
Procedural Generation	Experimental scripts
Dungeon Loader	Reads chunk configs

---

Web/AAC​

Feature	Status	Notes
Website Login	Working
Character Search	AJAX-powered
Highscores	Live leaderboard
Online List	Dynamic list
Admin Tools	Coming soon

---

Coming Soon (In Progress)​

Feature	Description
Instanced Dungeons	Scalable, configurable, group/solo
Advanced Monster AI	Hostility flags, memory, anger, element logic
Unity Client	Custom modern client (OTClient supported for now)
Plug-in System	Drop folder → auto-load scripts
Profession System	Forging, alchemy, crafting trees
Visual Overlays	Rarity glows, condition indicators
Quest System	Fully modular, flag/storage-based
Admin Dashboard	Live server stats & player tracking

---

Why Python?​

Benefit	How it Helps
Readable Syntax	Easier for new devs to contribute
Live Reloading	Change logic mid-game
Built-in Libs	File handling, math, data, etc.
No Compiling	Fast testing + fewer crash risks
Modular Code	Systems live in separate folders
Community Friendly	Great for teaching/gamejam servers

---

How to Get Started​

1. 
   
2. Clone the repo
   git clone https://github.com/Mystikil/PyOT
3. 
   
4. Install Python 3.11+ and dependencies
   pip install -r requirements.txt
5. 
   
6. Setup your MySQL database using the schema provided
7. 
   
8. Run the login server:
   python loginserver.py
9. 
   
10. Run the game server:
    python gameserver.py
11. 
    
12. (Optional) Launch the website at /website/ in any PHP server or XAMPP

---

Want to Contribute or Watch?​

If you:

• 
  
• Know Python
• 
  
• Are tired of hacking in Lua/C++
• 
  
• Love MMO or OT design
• 
  
• Want to build new systems cleanly
• 
  
• Just want to see something different succeed…

We’re building this out in the open.
Check out the repo:
GitHub - Mystikil/PyOT (https://github.com/Mystikil/PyOT)

---

Future Plans: PyOTV2​

We’ll be renaming the project to PyOTV2 soon.
That version will include:

• 
  
• Unity client integration
• 
  
• Full plug-and-play scripting
• 
  
• New AI core ("ECHO")
• 
  
• Admin panel
• 
  
• Secure serialization + login security

 

[~X~]

Will be updating this system to use the communities open sprites pack.

 

[Shawak]

Did you actually test how many players this server supports? I can imagine python being single core could be a problem relatively fast

 

[OTAmator]

Amazing work

 

[~X~]

Did you actually test how many players this server supports? I can imagine python being single core could be a problem relatively fast

This would all depend on the actual hardware. However Stain previously mentioned should withstand a great number of players. I personally have not tried it but I see no issue housing 1k players at once. But I run all industrial grade equipment.

Amazing work

Thank you!

 

[Shawak]

Well you are saying that but doing pathfinding in python and smth like this:

# A cache, this can probably get pretty big, but right now it's not something I'll think about
RouteCache = {} # {(FromX, FromY, ToZ, ToY, Z): [Route]}

doesn't really sounds promising

 

[~X~]

Well you are saying that but doing pathfinding in python and smth like this:

# A cache, this can probably get pretty big, but right now it's not something I'll think about
RouteCache = {} # {(FromX, FromY, ToZ, ToY, Z): [Route]}

doesn't really sounds promising

Ive yet to run into a path finding issues, but path finding is something that is being reworked via play time. This means its monitored, logged, and checked.. Then once done its analyzed and if needed things change. Also worth noting — route caching is dangerous if done naively. That's why I'm leaning toward scoped or LRU-based caching for hot zones or repeated paths (like temple exits or instance bosses). Not global A→B maps.

 

[El Diablo]

I'm slowly learning coding on my own time after work using LUA for personal hobby scripts and also taking boot.dev courses which has a focus on python. I saw this and was delighted. In the github though it does state 3.4.1 for the python version. Will this work with the latest version of python and tornado library?

This could be a valuable and fun way for me to continue learning coding!

 

[Ezzz]

Have you benchmarked this with at least 3000 active monsters ?

 

[jeanphilip]

with this Tibia 7.7 Server Decompiled (https://otland.net/threads/tibia-7-7-server-decompiled.296811/#post-2782789) and this project think we could make a huge improvement or base-pack for this project

 

[Source]

Last time I checked the original PyOT project out many years ago out, it did not stand basic scrutiny. I didn't get to test much out, all I remember was monsters were broken, stopped walking, IIRC suddenly glitched around, etc, so I just logged out and deleted the files off my computer.
Not to mention the code was unreadable, imo worse than OTServ C++ code.

The lack of any good alternatives to the massively bloated OTServ based projects is why I started my own project (TaoEngine) from scratch many years ago myself, so I can sort of understand the wish to work on something like this.

Instead of an LLM generated thread, I would've liked to see what this thing can do, and whether it stands basic scrutiny, in something like a youtube video, I think that would be pretty cool and a fair request, and without 3 second cherry picked clips where things doesn't break - let us see it break, and show us the progress along the way getting this thing not to break That's my opinion.

I'd like to be positive and hopeful, but it's kind of hard without any demonstrations of what it can (or can't) actually do yet.
That said, I wish you the best, hope you can show us more of what this can do in the future! Good luck with the project!

 

[~X~]

Last time I checked the original PyOT project out many years ago out, it did not stand basic scrutiny. I didn't get to test much out, all I remember was monsters were broken, stopped walking, IIRC suddenly glitched around, etc, so I just logged out and deleted the files off my computer.
Not to mention the code was unreadable, imo worse than OTServ C++ code.

The lack of any good alternatives to the massively bloated OTServ based projects is why I started my own project (TaoEngine) from scratch many years ago myself, so I can sort of understand the wish to work on something like this.

Instead of an LLM generated thread, I would've liked to see what this thing can do, and whether it stands basic scrutiny, in something like a youtube video, I think that would be pretty cool and a fair request, and without 3 second cherry picked clips where things doesn't break - let us see it break, and show us the progress along the way getting this thing not to break That's my opinion.

I'd like to be positive and hopeful, but it's kind of hard without any demonstrations of what it can (or can't) actually do yet.
That said, I wish you the best, hope you can show us more of what this can do in the future! Good luck with the project!

You are completely write! And the post was generated by the AI I built which works with me on the server. Soon as I get back I will update things as you mentioned!!!

Post automatically merged: Aug 22, 2025

Have you benchmarked this with at least 3000 active monsters ?

I have ran the server with a full RL Tibia map and no issues have came up with monster's. The only issues I did see was of my own which was a direct result of my instances system.

 

[Ezzz]

You are completely write! And the post was generated by the AI I built which works with me on the server. Soon as I get back I will update things as you mentioned!!!

Post automatically merged: Aug 22, 2025

I have ran the server with a full RL Tibia map and no issues have came up with monster's. The only issues I did see was of my own which was a direct result of my instances system.

Idle monsters don't count to measure how the server will behave.

 

[~X~]

Idle monsters don't count to measure how the server will behave.

I can arrange a large number if needed. That is not a issue!

 

[Ezzz]

I can arrange a large number if needed. That is not a issue!

I'm eager to see performance! From a programmer's perspective, I want to see how Python interacts with at least 1000 active monsters walking around and casting spells, I want to see if it would throttle or remain good,

 

[~X~]

Updated the site to be more modern, updated with information, and added a download button to download the current release. Please check out all the new information and features! Visit the site at PyOT Overview — PyOT Developer Guide (https://devnexus.online/pyot)

Post automatically merged: Oct 13, 2025

Last time I checked the original PyOT project out many years ago out, it did not stand basic scrutiny. I didn't get to test much out, all I remember was monsters were broken, stopped walking, IIRC suddenly glitched around, etc, so I just logged out and deleted the files off my computer.
Not to mention the code was unreadable, imo worse than OTServ C++ code.

The lack of any good alternatives to the massively bloated OTServ based projects is why I started my own project (TaoEngine) from scratch many years ago myself, so I can sort of understand the wish to work on something like this.

Instead of an LLM generated thread, I would've liked to see what this thing can do, and whether it stands basic scrutiny, in something like a youtube video, I think that would be pretty cool and a fair request, and without 3 second cherry picked clips where things doesn't break - let us see it break, and show us the progress along the way getting this thing not to break That's my opinion.

I'd like to be positive and hopeful, but it's kind of hard without any demonstrations of what it can (or can't) actually do yet.
That said, I wish you the best, hope you can show us more of what this can do in the future! Good luck with the project!

Your more than welcome to test out the server! Sorry Im super late replying but I had emergency shoulder surgery and it set me back a bit, but I'm back at it!

 

[~X~]

Tying a couple of the discussions together so please use this new thread located at Devnexus Online (https://otland.net/threads/devnexus-online.293923/) you can find updated information about the PyOT project there!