- Joined
- Nov 23, 2008
- Messages
- 2,050
- Reaction score
- 256
Introduction
I know discussions about the original CipSoft server have been around for many years. Some people have moved on, others are more interested in newer projects, and that's completely understandable. This thread isn't really aimed at them. It's for the people who still appreciate Tibia 7.4, the ones who have spent years trying to recreate that experience, and the ones who still dream about hosting an authentic server one day.
Even after more than twenty years, there is still an incredible amount of interest in 7.4. New servers continue to launch, players continue to return, and developers continue to invest thousands of hours trying to recreate something that was never truly preserved. I don't think that interest exists because people simply want another Open Tibia distribution. I think it exists because the original CipSoft server represents an important part of Tibia's history, and many people still want to experience or preserve it as accurately as possible. The Lazarus Project was built for those people.
The Lazarus Project aka TLP
Past couple of months I've been working on what I call The Lazarus Project. It's a reconstruction of
the 7.7 Cipsoft Binary from scratch made by me and downgraded to 7.4
The goal was never to build another Open Tibia server. It wasn't to make another custom distribution either. The goal was to reconstruct the original 7.4 server as faithfully as possible using the original CipSoft binary as the foundation for
the entire project. Everything in Lazarus Project has been built around one principle. Accuracy. Not assumptions. Not approximations. Not "this is probably how it worked." We wanted to know how the original server actually behaved, and then rebuild it based on evidence.
I know discussions about the original CipSoft server have been around for many years. Some people have moved on, others are more interested in newer projects, and that's completely understandable. This thread isn't really aimed at them. It's for the people who still appreciate Tibia 7.4, the ones who have spent years trying to recreate that experience, and the ones who still dream about hosting an authentic server one day.
Even after more than twenty years, there is still an incredible amount of interest in 7.4. New servers continue to launch, players continue to return, and developers continue to invest thousands of hours trying to recreate something that was never truly preserved. I don't think that interest exists because people simply want another Open Tibia distribution. I think it exists because the original CipSoft server represents an important part of Tibia's history, and many people still want to experience or preserve it as accurately as possible. The Lazarus Project was built for those people.
The Lazarus Project aka TLP
Past couple of months I've been working on what I call The Lazarus Project. It's a reconstruction of
the 7.7 Cipsoft Binary from scratch made by me and downgraded to 7.4
The goal was never to build another Open Tibia server. It wasn't to make another custom distribution either. The goal was to reconstruct the original 7.4 server as faithfully as possible using the original CipSoft binary as the foundation for
the entire project. Everything in Lazarus Project has been built around one principle. Accuracy. Not assumptions. Not approximations. Not "this is probably how it worked." We wanted to know how the original server actually behaved, and then rebuild it based on evidence.
The binary became the specification
The binary became the specification
Recovering the source code wasn't the difficult part and alot of people have already done it.
The difficult part was proving that the recovered implementation actually behaved like the original executable.
Because of that, every important subsystem has been verified in three different ways before being accepted.
1# Static verification against the original binary.
2# Runtime verification by observing the original executable while it was running.
3# Behavioral verification by executing identical scenarios against both implementations and comparing the results.
The original executable wasn't simply used to recover source code. It became the specification that every important subsystem had to satisfy. This philosophy ended up influencing every decision made throughout the project.
The difficult part was proving that the recovered implementation actually behaved like the original executable.
Because of that, every important subsystem has been verified in three different ways before being accepted.
1# Static verification against the original binary.
2# Runtime verification by observing the original executable while it was running.
3# Behavioral verification by executing identical scenarios against both implementations and comparing the results.
The original executable wasn't simply used to recover source code. It became the specification that every important subsystem had to satisfy. This philosophy ended up influencing every decision made throughout the project.
More than a decompilation
Over the years we've seen reverse engineering efforts, including Fusion's 7.7 reconstruction. Lazarus Project has a different objective. The goal has never been to recover source code that looks correct. The goal has been to recover source code that has been continuously proven against the original executable.
Every important recovery has been challenged statically, at runtime and behaviorally before becoming part of the final engine. The result isn't just recovered source code. It's a reconstruction backed by evidence.
Every important recovery has been challenged statically, at runtime and behaviorally before becoming part of the final engine. The result isn't just recovered source code. It's a reconstruction backed by evidence.
Reconstructing the original game
Recovering the engine was only one part of the project. The world itself also mattered. The Lazarus Project is based on the original 7.5 map that existed in Toors leaked 7.7 version. Those files have been cleaned, verified and reconstructed into a proper original map instead of relying on years of community modifications and conversions. This is the closest the community will get to the actual 7.4 map as it has the same borders and boats as the 7.4 map.
On top of that the reconstructed sources and files have been systematically brought back to a true 7.4 state. This wasn't simply a matter of changing protocol versions or replacing the map. It was in itself months of work prior to this reverse engineering began.
A large amount of work went into researching and restoring gameplay mechanics, NPC behaviour, movement rules, travel systems, progression, protocol differences, world interactions and numerous details that changed between 7.4 and later versions.
Rather than following how community distributions implemented these systems over the years, every important change was researched and restored with historical accuracy as the primary objective.
On top of that the reconstructed sources and files have been systematically brought back to a true 7.4 state. This wasn't simply a matter of changing protocol versions or replacing the map. It was in itself months of work prior to this reverse engineering began.
A large amount of work went into researching and restoring gameplay mechanics, NPC behaviour, movement rules, travel systems, progression, protocol differences, world interactions and numerous details that changed between 7.4 and later versions.
Rather than following how community distributions implemented these systems over the years, every important change was researched and restored with historical accuracy as the primary objective.
Modern AI meets reverse engineering
One thing I also think is worth sharing is how this project was built.
Almost the entire reconstruction has been performed using GPT-5.5 as the primary reverse engineering assistant. Over the course of roughly three months and approximately $2,200 in AI usage, it has helped analyze thousands of functions, reconstruct complex subsystems, debug recovered implementations and continuously compare them against the original engine.
The important part wasn't asking AI to generate code. The important part was building a workflow where every recovery had to be supported by evidence.
Recover -> Verify -> Compare -> Accept -> Repeat
That workflow has ultimately become the foundation of Lazarus Project. I honestly believe this project demonstrates how powerful modern AI can become when it's used as an engineering tool instead of a code generator.
What other changes were made?
Recovering the engine was only one part of the project. Restoring an authentic 7.4 experience required a significant amount of historical research and manual verification across the entire game.
Many of the original 7.4 mechanics had to be restored, including systems such as overspawn, healing while standing on stacked creatures, removal of despawn, range, fishing, regeneration, fluid formulas, magic level progression and numerous gameplay details that changed in later versions.
Around 67 NPCs were individually reviewed and a bunch of monster loot was updated to match their original 7.4 behavior. Boats and carpets once again function without protection zone restrictions, soul points were removed, quests and world interactions were revalidated, and the integrity of the original map was verified to ensure that moveuse actions and interactions behaved correctly A bunch of items had to be fixed in objects.srv aswell to accurately work well with the new 7.4 version such as the tile borders.
Those are only some of the larger changes. Countless smaller details were researched and restored throughout the project, many of which I dont even remember at this point but collectively define how the original game actually felt to play.
By the end of the project, the challenge was no longer recovering major systems. It was making sure the thousands of small details matched the original experience as closely as possible.
Almost the entire reconstruction has been performed using GPT-5.5 as the primary reverse engineering assistant. Over the course of roughly three months and approximately $2,200 in AI usage, it has helped analyze thousands of functions, reconstruct complex subsystems, debug recovered implementations and continuously compare them against the original engine.
The important part wasn't asking AI to generate code. The important part was building a workflow where every recovery had to be supported by evidence.
Recover -> Verify -> Compare -> Accept -> Repeat
That workflow has ultimately become the foundation of Lazarus Project. I honestly believe this project demonstrates how powerful modern AI can become when it's used as an engineering tool instead of a code generator.
What other changes were made?
Recovering the engine was only one part of the project. Restoring an authentic 7.4 experience required a significant amount of historical research and manual verification across the entire game.
Many of the original 7.4 mechanics had to be restored, including systems such as overspawn, healing while standing on stacked creatures, removal of despawn, range, fishing, regeneration, fluid formulas, magic level progression and numerous gameplay details that changed in later versions.
Around 67 NPCs were individually reviewed and a bunch of monster loot was updated to match their original 7.4 behavior. Boats and carpets once again function without protection zone restrictions, soul points were removed, quests and world interactions were revalidated, and the integrity of the original map was verified to ensure that moveuse actions and interactions behaved correctly A bunch of items had to be fixed in objects.srv aswell to accurately work well with the new 7.4 version such as the tile borders.
Those are only some of the larger changes. Countless smaller details were researched and restored throughout the project, many of which I dont even remember at this point but collectively define how the original game actually felt to play.
By the end of the project, the challenge was no longer recovering major systems. It was making sure the thousands of small details matched the original experience as closely as possible.
Final Thoughts
I recently made a comment that I had helped two guys downgrade Fusions to 7.4. But as off the completion of my own project, I've requested them to instead use this engine. I feel more confident with it in production and we will see it with players in a near future.
The Lazarus Project was built and reconstructed from scratch using every practical source of evidence available from the original executable. The objective wasn't simply to recover source code, but to recover and verify every observable aspect of the original engine that could be established with confidence.
Static analysis formed the foundation of the project, but it was never considered sufficient on its own. It provided the structure of the original server, while runtime analysis supplied evidence that could not always be established reliably from static data alone. Throughout the project, runtime analysis revealed execution paths, parser behavior, cleanup sequences, protocol state transitions and numerous implementation details that were either ambiguous or impossible to confirm statically. Those observations became part of the reconstruction instead of relying on interpretation.
The final stage was behavioral verification. Every important subsystem was executed under identical scenarios against both the original executable and the recovered implementation. This allowed subtle behavioral differences to be identified, investigated and corrected before becoming part of the final engine.
Lazarus Project was designed with production use as the end goal. That objective influenced every engineering decision throughout the reconstruction. Accuracy wasn't only about recovering source code, it was about reaching a level of confidence where the recovered implementation could realistically be deployed.
Another important aspect of The Lazarus Project is that the verification process was performed before the recovered codebase was cleaned or refactored. Static recovery, runtime analysis and behavioral verification were all carried out while the reconstruction still closely reflected the original binary. Only after a subsystem had been verified was it cleaned and integrated into the production codebase. This preserved a direct link between the recovered implementation and the original executable throughout the reconstruction process.
The workflow itself wasn't designed to trust either the AI or the engineer. Every important recovery had to be supported by evidence before it became part of the project. That process dramatically reduced subjective interpretation and guesswork throughout the reconstruction.
I chose GPT-5.5 because, after extensive experimentation, it consistently performed best for the scale and complexity of this project. Its ability to maintain context across a reconstruction of this size dramatically accelerated development while still requiring every important recovery to be verified against the original executable before being accepted.
Without modern AI, I genuinely believe a project of this scope would have required many years of work to reach the same level of verification. GPT-5.5 compressed what I believe would have been a decade-scale reverse engineering effort into roughly three months of focused reconstruction and validation.
Lazarus Project is the result of combining three independent sources of evidence into a single reconstruction. Static analysis established the structure, runtime analysis recovered behavior that static analysis alone could not fully establish, and behavioral verification confirmed that the recovered implementation behaved like the original executable in practice.
That philosophy defined every stage of the project and ultimately became the reason Lazarus Project was built for production rather than simply for research.
At last this is not a fake promise of a china version 7.4 server . This is the inner essence of that Cipsoft Server.
In every aspect it behaves identically and it has been proven.
The Lazarus Project was built and reconstructed from scratch using every practical source of evidence available from the original executable. The objective wasn't simply to recover source code, but to recover and verify every observable aspect of the original engine that could be established with confidence.
Static analysis formed the foundation of the project, but it was never considered sufficient on its own. It provided the structure of the original server, while runtime analysis supplied evidence that could not always be established reliably from static data alone. Throughout the project, runtime analysis revealed execution paths, parser behavior, cleanup sequences, protocol state transitions and numerous implementation details that were either ambiguous or impossible to confirm statically. Those observations became part of the reconstruction instead of relying on interpretation.
The final stage was behavioral verification. Every important subsystem was executed under identical scenarios against both the original executable and the recovered implementation. This allowed subtle behavioral differences to be identified, investigated and corrected before becoming part of the final engine.
Lazarus Project was designed with production use as the end goal. That objective influenced every engineering decision throughout the reconstruction. Accuracy wasn't only about recovering source code, it was about reaching a level of confidence where the recovered implementation could realistically be deployed.
Another important aspect of The Lazarus Project is that the verification process was performed before the recovered codebase was cleaned or refactored. Static recovery, runtime analysis and behavioral verification were all carried out while the reconstruction still closely reflected the original binary. Only after a subsystem had been verified was it cleaned and integrated into the production codebase. This preserved a direct link between the recovered implementation and the original executable throughout the reconstruction process.
The workflow itself wasn't designed to trust either the AI or the engineer. Every important recovery had to be supported by evidence before it became part of the project. That process dramatically reduced subjective interpretation and guesswork throughout the reconstruction.
I chose GPT-5.5 because, after extensive experimentation, it consistently performed best for the scale and complexity of this project. Its ability to maintain context across a reconstruction of this size dramatically accelerated development while still requiring every important recovery to be verified against the original executable before being accepted.
Without modern AI, I genuinely believe a project of this scope would have required many years of work to reach the same level of verification. GPT-5.5 compressed what I believe would have been a decade-scale reverse engineering effort into roughly three months of focused reconstruction and validation.
Lazarus Project is the result of combining three independent sources of evidence into a single reconstruction. Static analysis established the structure, runtime analysis recovered behavior that static analysis alone could not fully establish, and behavioral verification confirmed that the recovered implementation behaved like the original executable in practice.
That philosophy defined every stage of the project and ultimately became the reason Lazarus Project was built for production rather than simply for research.
At last this is not a fake promise of a china version 7.4 server . This is the inner essence of that Cipsoft Server.
In every aspect it behaves identically and it has been proven.
Why I'm not releasing it publicly
The Lazarus Project includes a production-ready ecosystem consisting of the game server, website, login manager, query manager, database schema and client. The Database is MySQL. With that said I do not want to see it become a part of a repository that is endlessly copied.
I previously released TVP 7.7 and TVP 7.4 with the hope of bringing the community together around a single 7.4 Open Tibia distribution that everyone could improve collectively. Unfortunately, that wasn't the outcome. Much of the work ended up being copied, repackaged and redistributed, while relatively little was contributed back to the original project.
Because of that, I've decided to take a different approach with The Lazarus Project. Rather than making it publicly available, it will be shared with a limited number of people who genuinely want to host an authentic 7.4 server and appreciate the work that went into reconstructing it. True enthusiasts. It's not for those interested in "researching it" or those who just want to "browse the files". This time I'm sorry but the community brought this on themselves.
I'd rather see the project in the hands of people who value it and use it for what I intended it to be rather than watching it become another repository that's endlessly copied.
So how do we get a copy?
My interest is to see more authentic 7.4 servers go live and have players. We can discuss your intentions, how you plan to advertise and maintaining
the server in the long run. I mainly want to filter out those who have no intentions of hosting and just want my files so that they can gather dust.
I previously released TVP 7.7 and TVP 7.4 with the hope of bringing the community together around a single 7.4 Open Tibia distribution that everyone could improve collectively. Unfortunately, that wasn't the outcome. Much of the work ended up being copied, repackaged and redistributed, while relatively little was contributed back to the original project.
Because of that, I've decided to take a different approach with The Lazarus Project. Rather than making it publicly available, it will be shared with a limited number of people who genuinely want to host an authentic 7.4 server and appreciate the work that went into reconstructing it. True enthusiasts. It's not for those interested in "researching it" or those who just want to "browse the files". This time I'm sorry but the community brought this on themselves.
I'd rather see the project in the hands of people who value it and use it for what I intended it to be rather than watching it become another repository that's endlessly copied.
So how do we get a copy?
My interest is to see more authentic 7.4 servers go live and have players. We can discuss your intentions, how you plan to advertise and maintaining
the server in the long run. I mainly want to filter out those who have no intentions of hosting and just want my files so that they can gather dust.
Last edited:
