#The AI isn't responding.

29 messages · Page 1 of 1 (latest)

magic rain
#

Hi! This post is regarding an issue I've been having for the past few weeks. I am using the 'Voices of the Court' mod for Crusader Kings 3, and, despite having done nothing beforehand, and the app/mod working completely fine a few days ago, the AI has just stopped responding. I know the AI can take a while to respond, sometimes up to a minute, but truly, I'll type out something, and nothing happens. I am pretty certain this is unlikely to be an issue with the mod itself, but maybe the server, and I'd be happy if you could aid me here.

Note: the connection is usually stated as valid, though sometimes there will be an error speaking of 'High Demand' for a particular model I am using. Also, I am using a free model, not a paid one.

viral marshBOT
#

I'm an AI assistant - my answers may be incorrect and I cannot see screenshots. Please verify information with the OpenRouter docs.

To help debug your issue, I need a few details:

  1. What is the exact error message or response you see when the AI does not respond (if any at all)?
  2. Which specific free model are you using (e.g., shisa-ai/shisa-v2-llama3.3-70b:free, nousresearch/deephermes-3-llama-3-8b-preview:free, etc.)?
  3. What is the relevant code snippet, API call, or steps you use to invoke the AI in the mod?

These details will help me (and the mods if escalation is needed) pinpoint the cause. Also, if you see “High Demand” messages, it may indicate the free model’s capacity is temporarily maxed out. For the most current status, check the OpenRouter status page or the #announcements channel in Discord.

-# This is an automated response. Use !mute to stop the bot from responding.

OpenRouter
OpenRouter

The unified interface for LLMs. Find the best models & prices for your prompts

OpenRouter
OpenRouter Status

OpenRouter Incident History

magic rain
# viral marsh *I'm an AI assistant - my answers may be incorrect and I cannot see screenshots....
  1. There usually isn't an error, save for the 'High Demand' messages.
  2. I am using google/gemini-2.0-flash-exp:free at the moment, but the issue persists even when using meta-llama/llama-3.3-70b-instruct:free. This, usually doesn't happen.
  3. Well, mostly I install the Voices of the Court app, then I establish the folder path. Afterwards, if I apply the mod itself on the game's launcher, everything usually works fine.
viral marshBOT
#

Thank you for the details. Since this issue affects multiple free models (such as google/gemini-2.0-flash-exp:free and meta-llama/llama-3.3-70b-instruct:free) and you usually don't see any error, a few points to consider:

  1. Free Model Limitations: Free variants like google/gemini-2.0-flash-exp:free may experience heavy demand and rate limiting, which can result in requests not being served or significant delays. The documentation states these models are subject to "rate-limiting and may be redirected or removed in the future" Gemini 2.0 Flash Experimental.
  2. No Error Response: If no error is returned and the connection appears valid, it may indicate your request is queued or silently dropped due to load or quota limits for free users. This is especially likely during peak times.
  3. Model/Mod Status: There are no known deprecations or removals for meta-llama/llama-3.3-70b-instruct:free at this time, and both models are still listed as available Llama 3.3 70B Instruct, Google Vertex Models.

Next steps:

  • Test with other free models (e.g., nousresearch/deephermes-3-llama-3-8b-preview:free or cognitivecomputations/dolphin3.0-mistral-24b:free) to see if the problem persists.
  • If you experience issues across all free models and still get valid connections but no replies, it's likely a temporary rate-limiting or quota issue.
  • For real-time updates on any platform-wide model outages or throttling, check the OpenRouter status page and #announcements channel in Discord.

If the issue continues despite these steps, please contact the mods through the Discord server with the details you’ve provided so they can investigate further.

-# This is an automated response. Use !mute to stop the bot from responding.

Gemini 2.0 Flash Experimental - API, Providers, Stats

Gemini Flash 2.0 offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1. Run Gemini 2.0 Flash Experimental with API

OpenRouter
OpenRouter

The unified interface for LLMs. Find the best models & prices for your prompts

OpenRouter
OpenRouter

The unified interface for LLMs. Find the best models & prices for your prompts

OpenRouter
OpenRouter Status

OpenRouter Incident History

magic rain
viral marshBOT
#

Bot has reached the maximum number of messages for this thread.

magic rain
#

<@&1384697330254610442> I hate to @ you lot, and I know you likely would rather do something else today, but I would ask for some assistance.

magic rain
#

Update: The problem persists when I do use the other free models suggested.

#

I don't know how 'temporary' this debacle is, considering its been happening for weeks.

magic rain
#

<@&1384697330254610442> @smoky lagoon

surreal sky
#

Ping me when you respond btw

magic rain
surreal sky
magic rain
#

No, I have not, I'll check to see if that works.

#

First I am unsubscribing and resubscribing to the mod on Steam to see if that's the issue, if it doesn't work, I'll go through the hassle of deleting and reinstalling the app itself.

#

It seems to be an app-based problem as the AI still does not respond.

magic rain
surreal sky
magic rain
surreal sky
magic rain
#

Well, I determined that I failed to delete it properly. I just did and have reinstalled it. The problem is, is now having to wait for the credits to return because I've already exceeded the rate limit or something.

surreal sky
#

[Key Redacted]

#

I'll deactivate the key once you've tested it

magic rain
#

Okay.

#

It works!

surreal sky
#

I deleted the key