This is an AzerothCore module built around mod-playerbots, breathing new life into bots by turning them into true in-game characters — companions with pre-defined personalities, memory, and relationships. Heavily inspired by mod-ollama-chat, but taking a different, more complex approach — focusing on the roleplaying experience rather than emulating real WoW players.
Think old Bioware games with companions — that's the core idea. The intended use is a fresh start at low rates with a full party of altbots playing alongside you, developing their own stories as you progress through the game together.
Important
The module is currently in active development and things could be changing rapidly. Before updating your copy, it's highly recommended to do a database backup, read the changes and notes in the Releases and check your server logs after running the newer version. There should be no hard incompatibilities, but new config variables are getting added, existing ones change their defaults, old routines get replaced and so on.
This module extends the bots provided by mod-playerbots into full characters — giving them a voice, memories, relationships, and the ability to react to events and converse with companions. The bot logic (combat, movement, questing) remains entirely untouched; this module only adds the personality layer on top.
Characters react to in-game events (chat, item pickups, duels, level-ups, boss kills, quests) based on configurable reply chances. When a character rolls to respond, a prompt is built from its character card, chat history, relationships, live context, and the event itself — then sent to a compatible LLM API, and the response is spoken by the character in-game.
Over time, chat history grows. When it reaches the configured token limit, a condensation process extracts discrete narrator-style memories with importance scores from the history, then clears all history. These memories are selected by importance at prompt-build time within a token budget, keeping the in-memory context bounded while preserving what matters. This way, characters gradually develop lasting memories and personality traits.
Relationships are tracked between characters and real players, as well as other characters. When a name is mentioned enough times in a character's history, a relationship update is triggered — generating or updating a brief description of how the character feels about that person. These are included in future prompts, giving characters continuity with their companions.
The module supports multiple languages based on the server's DBC.Locale setting:
- enUS (default, picked if the needed locale is not found)
- deDE
- ruRU
Since mod-playerbots is an obvious hard requirement, follow their Installation Guide until you have a working acore installation with playerbots enabled.
There are three ways to obtain the module sources:
Clone a specific tagged release:
cd modules
git clone --depth 1 --branch <tag_name> [email protected]:deseven/mod-playerbots-characters.gitWhen you want to upgrade to a newer release tag:
cd modules/mod-playerbots-characters
git fetch --tags
git checkout <new_tag_name>A list of all available tags can be found on the Releases page.
Download a .zip or .tar.gz source archive from the Releases page and extract it into the modules directory. The extracted folder must be named mod-playerbots-characters.
Clone the repository normally to get the latest changes:
cd modules
git clone [email protected]:deseven/mod-playerbots-characters.gitThis gives you the most recent commits from the main branch. You can git pull at any time to update.
After obtaining the sources using any of the methods above, rebuild the server normally.
Note
The module includes bundled copies of nlohmann/json in deps/nlohmann/json.hpp and yhirose/cpp-httplib in deps/yhirose/cpp-httplib/httplib.h, so no external libraries are required.
If mod_weather_vibe is also installed, weather states from it will be included in the character's scene description.
Copy env/dist/etc/modules/playerbots_characters.conf.dist as env/dist/etc/modules/playerbots_characters.conf and adjust it as needed.
Each LLM connection is defined in a standalone JSONC file (JSON with comments). To get started, pick one of the ready-made .jsonc.example templates in the connections/ directory, copy it to a new file without the .example suffix, and point the relevant config parameter at it. For the full setup guide — supported API types, config parameters, tested providers and per-task routing — see Model Connections.
After configuring, use .chars connection-test (optionally with a connection name like utility) to verify that a connection is working.
Important
Due to the complexity and length of the prompts, locally-run models on average home hardware will generally struggle and produce low-quality output as context grows. A cloud-based model with a large context window is recommended. Make sure to also adjust PBC.MaxHistoryCtx accordingly — a good starting point is around 25% of the model's total context window. Aim for at least 32k in general, anything less could lead to poor efficiency of character relationship tracking and memory extraction.
The module can optionally run a built-in HTTP/WS server with a web frontend, providing an interface for managing the characters, as well as an API for external tools and integrations. The web interface supports viewing and editing chat history, relationships and memories. This is disabled by default and could be enabled in the config, read the HTTP SERVER section and follow the comments there.
Note
If you plan to use the web app for a considerable amount of time without touching the game, it's also recommended to set PreventAFKLogout to 2 in your worldserver.conf.
Recommended adjustments to the playerbots config (playerbots.conf), to make bots less talkative in order to not interfere with the new character logic:
| Setting | Value | Purpose |
|---|---|---|
AiPlayerbot.EnableBroadcasts |
0 | Disables loot / quest / kill broadcasts |
AiPlayerbot.RandomBotTalk |
0 | Disables random talking in say / yell / general channels |
AiPlayerbot.RandomBotEmote |
0 | Disables random emoting |
AiPlayerbot.RandomBotSuggestDungeons |
0 | Disables dungeon suggestions in chat |
AiPlayerbot.EnableGreet |
0 | Disables greeting when invited |
AiPlayerbot.GuildFeedback |
0 | Disables guild event chatting |
AiPlayerbot.RandomBotSayWithoutMaster |
0 | Disables bots talking without a master |
AiPlayerbot.RPWarningCooldown |
999999 | Increases delay between "missing reagents" messages and such |
AiPlayerbot.EnableAutoTradeOnItemMention |
0 | Disables automatic trades and item listings on item mention (from mod-playerbots@8caf37a) |
Start the server, set up some altbots for yourself or invite existing random bots, then write cards for them as you see fit. See characters/Example.card.txt for an example of how to write a character card. The character name in the filename must match the in-game character name exactly to be picked up. It is recommended to write cards in second person (using "you/your"), because default prompts and request structure follow this format.
Start playing, chat with your characters, discuss anything you like, build relationships and enjoy the game.
Note
Depending on the model you are using, your mileage may vary. Do regular backups with modules/mod-playerbots-characters/tools/pbc_backup.sh and adjust things as needed either in the database (followed by .chars reload command) or via the included web app. There are also two helper tools (pbc_info.sh and pbc_history.sh) which might help with tracking what's going on. You can also steer the narration a bit by using .chars narrate and .chars narrate-party commands. Check out available commands for more info.
If you have the HTTP server configured, you can use .chars web command to get a one-time password that you can then use to authorize in the web app.
Two config options control debug output:
PBC.DebugEnabled— enables general debug logging (event dispatching, roll chances, etc.)PBC.DebugShowFullRequest— when enabled alongsideDebugEnabled, logs the full request body and response body for every LLM API call. Useful for diagnosing issues with non-standard API providers where the response format may differ from OpenAI's.
- Model Connections — LLM connection setup, supported API types and tested providers
- Events — list of in-game events characters can react to
- Commands — available in-game and console commands
- Prompts — prompt templates, customization and template variables
- API Reference — HTTP API and WebSocket documentation
- Frontend — web interface build and development
Contributions are welcome — feel free to open a pull request. If you need help or found a bug, open an issue.