¶ FAQ
¶ When I start the CHIM server I get a “Webpage can not be found/HTTP 500” error!
Most likely the conf.php got corrupted somehow. Delete conf.php under /conf. Update the CHIM server and restart it. A new clean one will be generated.
¶ Why is the CHIM menu not showing in MCM?
You may be missing a mod requirement or dependency. Double check your downloaded and installed mods.
¶ I see an error message in the game that "AIAgent.ini file not present or invalid" and the Narrator/NPCs don't answer me?
1. Ensure that you have downloaded the AIAgent.ini file and placed it in the (CHIM)/SKSE/Plugins folder in your mod manager. See the CHIM Installation section for more details.
2. If you are using Vortex, click the "Deploy Mods" button after copying the ini file to the Plugins folder.
3. Check that you have all of the required mods listed on CHIM's Nexus page enabled.
4. Do not rename the AIAgent.ini file. The "ini" file extension may be hidden by default Windows settings. If you renamed the file, delete it, download it again, and place it in the Plugins folder.
¶ The game crashes after….
Most CTD’s are not a result of a CHIM error. You either have a mod that is incompatible (check the modpage) or another mod is causing a conflict elsewhere. Check your crashlogger for more info. If you are getting crash on newgame, then look further down this page for the appropriate FAQ article.
¶ Actions/Commands are not working!
Sometimes you may have to recheck the "Enable Actions" options in game in the MCM menu. Also make sure you are using a valid LLM that has JSON capabilities.
¶ How many AI can I have activated at once?
As many as you want, there is no hard coded limit! Also due to how the system works it is fine to keep AI activated when you are not in the same area. It will not use up any more resources compared to having a single AI activated. Resources are used when it generates a response.
¶ Do I have to restart the game if I make any changes in the CHIM server?
No, changes are applied instantly. You can easily make lots of changes in the configuration wizard while the game is running.
¶ The AI is not responding and I am getting an error message in the webserver and it looks like this:
Array ( [type] => 2 [message] => fopen(https://api.openai.com/v1/chat/completions): Failed to open stream: HTTP request failed! HTTP/1.1 401 Unauthorized [file] =>
A: There are 2 reasons for this occurring:
You are using a free account. You need to disable actions in game as you have a request limit of 3 API calls a minute.
Your API key has expired or your free trial has ended. You will either need to generate a new key or start a paid account.
¶ What do these "Valid Moods" mean for the Azure voice settings?
Azure TTS voices have options to speak in different tones (angry, sad, whispering etc) depending on the voice you picked. You can force the voice to speak ina specific way (e.g. only whisper). Here is their speech gallery with examples: https://speech.microsoft.com/portal/147f056015ad4f749492cde312faa1c6/voicegallery
¶ It's not working with my SteamDeck!
Make sure you have an etc/resolv.conf file. Then add these 2 lines to the file:
nameserver=8.8.8.8
nameserver=8.8.4.4
¶ How can I fully uninstall the DwemerDistro if I deleted the files before running the uninstall script?
Run this command in command prompt:
wsl --unregister DwemerAI4Skyrim3
¶ Will there be an LE port?
No
¶ Does this work with pirated Skyrim?
No
¶ CHIM causes my game to crash on newgame!
Some alternate start mods may have issues with AIAgent. We recommend:
Disabling AIAgent
Go through the newgame process with your mod
Save and exit once you get to "regular Skyrim"
Then reenable AIAgent.esp
¶ My custom uploaded voice file for CHIM XTTS is not working!
This is the recommend file specifications for uploading a voice to XTTS.
.wav format
PCM
16 bit
Mono
20500hz
¶ Is there a way I can move the Distro to another drive without having to uninstall and reinstall?
1. Shut down WSL (wsl --shutdown).
2. Copy the ext4.vhdx file from its original location to the new location.
3. Update the BasePath in the registry (under the distro’s GUID in HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Lxss) to point to the new folder.
4. Restart WSL and test.Most likely culprit is your machine is lagging trying to run too many AI services on top of heavily modded Skyrim! Try disabling any offline AI services to see if that frees up any VRAM. We do recommend using online services, such as LLM's, to run this mod.
¶ My game is running really slow and in the server error logs I keep seeing:
(PHP Fatal error: Uncaught Error: No PostgreSQL connection opened yet in)!
Most likely culprit is your machine is lagging trying to run too many AI services on top of heavily modded Skyrim! Try disabling any offline AI services to see if that frees up any VRAM. We do recommend using online services, such as LLM's, to run this mod.
¶ A lot of models don't work in Openrouter. Why is this?
First: try to use instruct models. Second: Everytime you change a model you must use the Autofill parameters button. Every model has different parameters. Try first using the "Autofill parameters low randomness" button.
We use the term "low randomness"as a simplification, but the truth is more complex than that. P10 means almost 90% of people are using these parameters set to work with this model. P90 will be a set of parameters only 10% people use.
Initial setup should use the P10. Once working, you can experiment with the parameters.
¶ Is OpenRouter free?
No. It's cheap, but not free. Has some free models, which we don't recommend to use because outputs can be cached (OR warns about this). They seem to allow some negative credit, but it will stop working at some point. Put some money into it, $5 is the minimal amount. Probably will last for a month playing every day. You can still use ChatGPT via OR.. Remember each follower/NPC can use a different LLM model..
¶ I loaded an old save and I lost a bunch of memories/events!
Due to how we track ingame time and events to sync with the server, we have to delete events that have happened in the "future". This is to prevent time paradoxes!
¶ vmmem/DwemerDistro3 is still running in the task manager and using up all my RAM!
Run the Stop.bat script, click Force Stop Server, or type this in the Command Prompt: wsl --shutdown
¶ What do I have to do to import [CUSTOM NPC]?
Just use the Upload NPC Biography tool in the webserver.
¶ XTTS is really slow and does not work!
Depending on your graphics card you may have difficulty running XTTS at a decent speed. It is a fantastic tool, however it requires a lot of power to run.
One thing you can do is CAP YOUR SKYRIM FRAME-RATE. As most Skyrim mod lists are highly unoptimized, capping your frame-rate can save you precious VRAM!
¶ Something is wrong and I have MinAI installed!
Read this: https://github.com/MinLL/MinAI/blob/main/FAQ.md
¶ I can hear the Narrator, but not any other NPC!
If you are using stereo sound, try disabling spatial audio. Otherwise play around with audio settings in Windows.
¶ The WSL/DwemerDistro server no longer works and exits out immediately when I launch it!
Here is the solution: https://github.com/microsoft/WSL/issues/8693#issuecomment-1319904954
¶ How to delete everything to start a new playthrough?
Any other settings will be cleared once it detects a new game.
¶ I want to see only subtitles and hear no audio from the AI NPC’s
Download and use the Mimic3 component. Then in the MCM CHIM audio to 0. Mimic3 will still “generate” an audio line, but you won’t hear it, and it will barely use your system resources. Lipsync code requires an audio file of some kind to be generated.
¶ ILLama Lava Bugs/Install Issues
Using a 4080/4090, the llama lava included in the distro fails to start with the above error
(see log in \\wsl.localhost\DwemerAI4Skyrim3\var\www\html\HerikaServer\log\vision.log).
To fix: in \\wsl.localhost\DwemerAI4Skyrim3\home\dwemer\llama-cpp edit the ddistro_install.sh file.
Change the line that starts with "make" to: make clean && make -j6 LLAMA_CUDA=1 LLAMA_CUBLAS=1 server
Then run Install Llama Lava (ITT Server).bat again.
¶ Can I use Groq?
Yes.
¶ AI Voices ingame are really quiet and sound weird when I have my back towards NPCs.
Most likely cause is how your Windows audio is configured.. Make sure you have stereo enabled and disable spatial audio. Also the MCM menu has volume sliders you can adjust.
¶ Can I use LLM Studio?
Yes, here is how:
Select the openrouterjson connector.
In the LM Studio developer tab, click settings and make sure Serve on Local Network is turned on.
Set ip address to http://(your ipv4 address):1234/v1/chat/completions
Ensure ENFORCE_JSON and json_schema are set to True.
Save changes and use the copy to all profiles buttons for each of the above settings.
¶ How to use a different GPU for CHIM XTTS
In case anyone else unfamiliar with python might need this in the future: to use solely a secondary gpu for xtts (if "--device cuda:1" isn't working); add the line "export CUDA_VISIBLE_DEVICES=1" as its own line in start-deepspeed.sh, before the line with "python -m". with line breaks above and below the line you've added, just like the lines already there. You can replace 1 with your gpu id, usually seen in task manager.
¶ Can I run XTTS on an AMD GPU
If you have an 7900GRE, 7900xt ,and 7900xtx then yes! Check out this:CHIM Manual