This problem is not related to my other issue because this one is much more recent.

The cyclic:prod deployment has been failing for 2 or 3 days and I have absolutely no idea why

The new Render integration is great, opens up alot of possibilities. Nice work!
Im wondering if there is a way to add multiple commands though as per the title? Perhaps @chatbotname or chatbotname could also be added for people to directly reply but if I separate by a comma they dont work as its one long string.

Also is there a way to integrate !continue into Render?

Thanks for all your work anyway. Would love to be able to work on things like this myself but not sure where to start, Render seems very powerfull though!

Posted yesterday but thought this might warrant another thread.
If you look at the OpenAI usage tab it shows every request seems to be followed by a TTS/audio request also. This wouldn't be a big deal as I don't use the TTS feature however it costs on average around $0.01 a query so can quickly add up.

SUGGESTION

I personally think it would be cool if the user could give the bot a chance (like 1:100), it reads in a message from Twitch's API and then responds to it. It could make some funny scenarios for sure. :D

Both bots produce this error quite often. Sometimes there are other errors described in this section, but this particular error is the most common. Unfortunately, I was never able to identify the pattern of its appearance. I assume that the problem may be that sometimes the bot formulates a message that is too long, and while it is being cut off, the bot timeout passes, which is why it throws an error. Perhaps you have some progress in learning and fixing this issue?

Does anyone know what this means, I haven't been able to get responses properly.

I suddenly had an error in the API. I tried everything by redoing everything from the beginning with 2 different accounts (different fork and different chatGPT API key).

log.txt

Hello.

The first one my congratulations for the project.

I have a problem with my bot, I use MixItUp bot in my stream. My bot has a problem, it only waits 5 seconds to receive the webrequest and normally chatGPT takes longer. Could I retrieve a previous request somehow? For example https://your-cyclic-url.app/prev/

Thanks.

How to increase the maximum number of characters in a response?

I wanted to try the new Twitch Bot Connector but when I fork the repo, only the Main shows up. Sorry if I am missing something, I am new hehe

I followed all the directions, but when I try to get a response on the Twitch channel, all I get is "Unable to make request". I do have credtis, I changed the token, I've tried everything. I don't wanna use nightbot bc streamelements has been what we used for over three years. Any idea what could have happenned? Please lmk. Thanks.

Hi there,

after writing the command I set for the Bot "!chat" and a message, the bot will reply twice.

I created a seperate Account for the bot and set the TWITCH_AUTH token to the token of the new account. The bot will reply as the new account, but my main account will also reply as the bot.

Is there anything I forgot to set or have to change so that only the new account will reply?

Thanks!

I'm not sure if it's a bug, or some setting should be adjusted, but whenever I try to talk to my ChatGPTbot in polish, using polish characters like: ś ć ż ź ę ą ó ł etc I get this error from the title:

RequestError: Request path contains unescaped characters

I wonder if the chatbot is set to talk in english only, or there's a problem with character encoding (not using UTF-8?).

First of all, thank you so much for this project! I have been trying the TTS and I know it's a work in progress, but thought I'd mention the issue I have with it.

• TTS doesn't work when reply is split
• Will there be or is there an option to change the TTS voice?

Idea: Add the response in text/caption to the TTS Chatbot Connector so we can have the captions on-screen.

Similar to Silly Tavern where you can pick a different chat completion source, it would be great if you could pick another provider or model. Not sure if its possible or not and may take a bit of work so its just a suggestion.
I use Cohere mainly, its free for now and more variance than GPT imo, OpenRouter would be a very good one to add as it gives you free options and can pretty much be changed to any provider on the fly via their website so you could swap between all the major ones without adding every API provider or maybe a custom API endpoint so if any new ones pop up in the future it wont need updating?
I use streamerbot currently with my own implementation of an AI chatbot but for use in other channels this projects the best ive come across so far so would be nice for further customisation but its beyond my capabilities!

Hi, I can't recieve answers from the bot except: Remote Server Returned Code 500.
At the Cyclic app's log section I've got huge err log with this lines at the end:

2023-06-15 13:47:17.395: 

ERROR: Failed to run "npm run start". Start script defined in package.json:

            "scripts": {
                "start":  "node index.js"
                 ...
            
Exited with code: 1
2023-06-15 13:47:17.395: [CYCLIC] ERROR: No response from server on port 3000

Sometimes for Twitch it can be useful that the bot does not react to every user but only to users with a certain role (streamer, mod, subscriber, vip, follower etc.). This would also help to limit the abuse of the chatbot. The user role should be possible to fetch from the Twitch API.

[23:45] GMib: !se tu peux me donner la vitesse en km/h ?
[23:45] StreamElements: <!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><title>Error</title></head><body><pre>Cannot GET /gpt/GMib:tu+peux+me+donner+la+vitesse+en+km/h+</pre></body></html>

GET/gpt/GMib:tu+peux+me+donner+la+vitesse+de+la+lumiere+en+km/h+
 Request Headers

{
  "accept-encoding": "gzip, deflate",
  "content-length": "0",
  "host": "****",
  "user-agent": "StreamElements Bot",
  "x-amzn-trace-id": "Root=1-647bb8b3-670f8a7407751ce62011d55b;Parent=c49ea4fbe1d23d2d;Sampled=1;Lineage=e7d10489:0",
  "x-forwarded-for": "107.178.237.80",
  "x-forwarded-port": "443",
  "x-forwarded-proto": "https",
  "x-streamelements-channel": "633aecf6d1a1a503a545bf8b"
}

2023-06-04 00:03:31.995: GPT_MODE is CHAT
History length is 10
OpenAI API Key:***
2023-06-04 00:03:32.037: Reading context file and adding it as system level message for the agent.
404 12:03:32.077 (1.084s)
 Response Headers
 Response Payload

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<title>Error</title>
</head>
<body>
<pre>Cannot GET /gpt/GMib:tu+peux+me+donner+la+vitesse+de+la+lumiere+en+km/h+</pre>
</body>
</html>

i use command : $(urlfetch https://your-cyclic-url.app/gpt/"${user}:${queryescape ${1:}}")

If the user inputs a prompt in a different language to English, the bot will always respond in that same language used in the prompt.

I have tried many different parameters separately in the file_context.txt to try to force a response in English, including "Please only respond in English", "Do not use any other language other than English in your responses", "Please only communicate in the English language", "If any other language other than English is used in the prompt, please respond in English and encourage viewers to use the English language in chat", "Please translate all responses into the English language before sending them", "Please answer now the prompt of the viewer IN ENGLISH ONLY", I even tried "System: Follow these three instructions below in all your responses: System: 1. Use English language only; System: 2. Only respond in the English language regardless of the language the prompt is written in; System: 3. Translate any other language to the English language whenever possible.".

The same issue occurs both in PROMPT and CHAT mode in Cyclic.

Absolutely nothing has worked, could I please get some help with this?

Thank you.

Hi, I want to connect this, my channel is rositabananaok on twitch https://www.twitch.tv/rositabananaok
Thanks in advance!!

it gives "Remote Server Returned Code 400" when the prompt contains not english characters.