Tried this on Firefox and Chrome, same thing happened.
When the text size is set to a small value (let's say up to 3-4) and there's already a lot of messages in the chat, changing the text size for a higher value triggers an autoscroll of the chat that can't be stopped. The more messages, the more it takes for the scroll to complete and scroll down until the most recent message.

hi it would be great to create a widget plugin for the avermedia live streamer ax310 as for elgato except that it brings more precisely

There is an exception in parsePlaceholders if the src argument is null.

It throws TypeError: can't access property "toString", n is undefined.

Also, in the above linked source, the toString() function is called twice in a row, I think the first call should be removed and a null check should be added.

I highly suggest to use an utility function like this instead of Object.toString():

function toStr(obj) { // prevent name clash even if toString should be safe
    if (obj === undefined) return null; // clean handling of undefined and null
    if (obj === null) return null;
    return Object.prototype.toString.call(obj); // ensure it has not been overriden to avoid any obscure bug
}

Hello I have obs 28.1.2 with integrated websocket but when I try to link twitchat to obs (to cut stream after raid) it says I need obs28+ OR websocket 5 when I try to install WS5 it doesn't work and says the plugin can't be loaded
Is it normal ?

Hello,

I have heard on a stream today a streamer that wanted to see on the chat as a "badge" how many streams a viewer has watched in a streak.

I directly thought to TwitchChat I think that could be something interesting.
I don't know if the badge it's the better option or instead maybe on the user infos of a viewer.

Btw, I love the job that has been done here !

Hello et bonjour !

Tout d'abord, un énorme GG pour cet outil incroyable ! du super taff, ultra pratique avec des fonctionnalités vraiment top !

Je me permet ce message ici car je ne sais ou le mettre d'autre.
J'ai un petit soucis avec les triggers :
J'ai un trigger sur une commande chat ( !kick )
J'ai 2 actions qui s'enchainent de contrôle d'OBS
La première, rendre une source active, cette source est un fichier media
J'ai activer l'option "attendre la fin du media"
La seconde, rendre une source inactive, cette même source

Sur ma scène principale gaming, c'est nickel. Sur toutes les autres, le trigger ne se déclenche pas.
Il a eu marché c'est sur (mes modos m'ont suffisamment fait sursauté avec ce son), mais depuis 3/4 semaines environ j'ai ce comportement. Je ne comprends pas. J'ai refait mes sources, refait le trigger, rien n'y fait. La source est bien sur présente sur toutes les scènes. J'ai même essayer d'imbriquer une scène dans une scène.
Si jamais vous avez une piste pour m'aiguiller ...

Autre chose, sur les triggers toujours, on a la possibilité de faire une condition, c'est top. Mais serait il possible de pouvoir imbriquer plusieurs conditions ? (en gros pouvoir remettre une condition sur le faux ^^ )
Je m'explique, toujours sur ma commande !kick, je souhaiterais que si c'est tel modo cela controle telle source (modo 1 source 1, modo 2 source 2 etc)

Encore une fois merci pour ce fabuleux outil !

Inbound whispers can already be shown in chat:

However, outbound whispers, e.g. the "/w " command, aren't shown in chat.
I personally whisper to viewers a bunch and while it does work, some feedback in chat would be good since I'm sometimes not sure whether I messaged the correct person.

If I create a trigger to move a source, it works only if "Animate change" is unchecked.

With "Animate change" checked, the source doesn't move, and here is the debug log from OBS:

info: [obs-websocket] [debug] [WebSocketServer::onMessage] Incoming message (decoded):
{
  "d": {
    "requestId": "484",
    "requestType": "GetCurrentProgramScene"
  },
  "op": 6
}
info: [obs-websocket] [debug] [WebSocketServer::onMessage] Outgoing message:
{
  "d": {
    "requestId": "484",
    "requestStatus": {
      "code": 100,
      "result": true
    },
    "requestType": "GetCurrentProgramScene",
    "responseData": {
      "currentProgramSceneName": "Start Scene"
    }
  },
  "op": 7
}
info: [obs-websocket] [debug] [WebSocketServer::onMessage] Incoming message (decoded):
{
  "d": {
    "requestData": {
      "sceneItemId": 53,
      "sceneName": "Start Scene"
    },
    "requestId": "485",
    "requestType": "GetSceneItemTransform"
  },
  "op": 6
}
info: [obs-websocket] [debug] [WebSocketServer::onMessage] Outgoing message:
{
  "d": {
    "requestId": "485",
    "requestStatus": {
      "code": 100,
      "result": true
    },
    "requestType": "GetSceneItemTransform",
    "responseData": {
      "sceneItemTransform": {
        "alignment": 5,
        "boundsAlignment": 0,
        "boundsHeight": 1.0,
        "boundsType": "OBS_BOUNDS_NONE",
        "boundsWidth": 1.0,
        "cropBottom": 0,
        "cropLeft": 0,
        "cropRight": 0,
        "cropTop": 0,
        "height": 84.0,
        "positionX": 58.0,
        "positionY": 100.0,
        "rotation": 0.0,
        "scaleX": 1.0,
        "scaleY": 1.0,
        "sourceHeight": 84.0,
        "sourceWidth": 1804.0,
        "width": 1804.0
      }
    }
  },
  "op": 7
}

There is no message to set the transform.

In the trigger log, I see:

4.451Animate transformation. Duration: undefined. Easing: sine.in

The animation duration was set to the default 500ms value. If I rewrite 500, then the animation works. It seems there is a problem with the default value in the form. I use firefox, and it might be why the behaviour is different.

As a side note, I see that the input is of type number. I highly suggest you do not use that.

• https://stackoverflow.blog/2022/12/26/why-the-number-input-is-the-worst-input/
• https://css-tricks.com/you-probably-dont-need-input-typenumber/
• https://css-tricks.com/what-to-use-instead-of-number-inputs/

Under linux, the font rendering is different and with the current settings, it is very thin (look at screenshot).

I think it would require quite some work to find some universal font setting. Instead I propose to add an advanced setting to specify font-weight and font-family css directly.

BTTV, FFZ and 7TV emotes are all case sensitive, which the browser addons respect. Twitchat doesn't care for case-sensitivity, which might cause emotes wrongfully being displayed, e.g. the sentence "many ppl do something" would have the ppL global 7TV emote inserted into that.

Hello,

Thanks for your work, i’m trying to self host the version, but the instruction to do it seems wrong.
In the Install server section you said : Next to this file, create a dist folder
But in the example structure

─ root
  ├─ node_modules/
  ├─ dist/
  ├─ controllers/
  ├─ utils/
  ├─ bootstrap.js
  ├─ env.conf
  ├─ credentials.json

the dist file is within the server folder, not next to it.
When the dist directory is in the server folder the directory can't be found.
But when the directory is next to the server folder its ok.

Beside this detail, when i start the server with node bootstrap.js there is no translation, only key of translation are visibles.

When the "pause on hove" option is enabled, the chat is nudged toward the top of about 20 pixels when hovering the chat.

When the text size is set to a value of 4 or higher, the minified badges are larger than the regular badges.
Plus I've noted in my case that I have 2 badges in the chat (broadcaster + verified), and when I activate the minified badges I have 3 badges, 1 purple (sub badge I presume) + 1 red (broadcaster) + 1 purple with a white dash (verified).

Hi 👋

My request would be to have the support for 7tv emotes.

7tv is an emote system like BTTV or FFZ that adds more emotes (base limit up to 150 emotes) and it's dynamically updated

https://7tv.app

On the trigger edition page, there are 4 buttons:

• Resync rewards
• Resync twitch extensions
• Test trigger
• Delete trigger

I suggest you add a toggle control to enable/disable the trigger from this page.

The use case is this: when I create a trigger "on the fly", for example "on any message", in a chat that is busy, I want to be able to test it by enabling it for like 5 seconds then turning it off, and this would be easier from the trigger page. Having to go back, find the trigger in the list, toggle it... is a bit of a usability problem.

I am currently trying twitchat to replace all my overlays, and I am missing one feature: custom overlays.

I would be great to have a custom overlay overlay, which would just be a text area for some raw code. It would be pretty low level and might be a little niche, but it would allow developers to create anything.

If you agree on the feature, I could work on it and make a PR.

A few features on the top of my head:

• custom html/css/svg/js
• access to all twitchat placeholders
• can be triggered/observe values

I have Twitchat always openon my secondary screen (because why not...)

But look at this screenshot of my spotify application dedicated to twitchat

It seems that the "currently-playing" endpoint is fetched very, very, VERY often. And If think it may causes my !song trigger to add songs to the queue to fail most of the time (nothing done even if the song exists)

PS: I use the hosted version of twitchat, I don't know if the API requests are sent by the front or backend

Hi there,

It seems like any other IP than the loopback ends with an attempt to open the OBS websocket over a secure channel (on wss://) resulting in not being able to control OBS from another computer on a local network (as it seems that OBS doesn't offer the ability to set a secure WebSocket). Would it be possible to offer the opportunity to make the secure access an opt-in option instead of a default behaviour ?

suggestion d'outil pour les streamers qui demandent à leurs viewers qui raider:

• les viewers tapent leurs suggestions de la façon : !raidprop SamuelEtienne
• l'interface twitchat fait automatiquement les stats des propositions des viewers,
• dans les propositions, on affiche aussi le nombre de viewers des gens suggérés et si le streamer le follow déjà. Ainsi que la date du dernier raid vers ce streamer (pour pouvoir alterner facilement)

Ainsi le streamer trouve plus facilement qui raider

Hello

Unless I am mistaken, I cannot find the chat overlay that can be placed in a browser source to display the chat to our viewers.

Does this feature still exist?

Thx for your work

Tried in differents tabs, works fine with Chrome but the scrollbar seems to be locked in Firefox when using the mouse. Using the Stream Deck plugin, I managed to scroll it up and down, but still not with the mouse even after that.

Windows 10 64-bit 21H2, Firefox 100.0

Hey :-) This project looks great. Just wondering how hard it would be to get this working with more than one twitch channel simultaneously? I have a special use case where I stream on two accounts simultaneously some of the time, kind of an experimental genre I'm trying out.

Twitch global and sub emotes are already in the emote autofill feature when typing ":" in chat:

It would be awesome if 3rd party emotes could also be quickly found with the autofill feature, e.g. typing "pp" would also show 7TV's ppL in the suggestions.

I started opening files in the project, and my editor is messing with the code automatically, especially it is triming trailing whitespace and ensuring newline at end of file, also it replaces tabs with spaces.

I suggest you add an .editorconfig file to the project to help with this.

I usually use trim_trailing_whitespace and insert_final_newline to both true, but you might set them to your liking, having the .editorconfig ensure editors of contributing developers will follow your style.

https://editorconfig.org/#supported-properties

I am trying to create a trigger that update the text of an OBS source, the source is a text source under linux but in the "Action" menu, I do not have the option to update the text.

If I use obs-websocket-js and list the items, my note item is as follow:

    {
      inputKind: 'text_ft2_source_v2',
      isGroup: null,
      sceneItemBlendMode: 'OBS_BLEND_NORMAL',
      sceneItemEnabled: true,
      sceneItemId: 40,
      sceneItemIndex: 14,
      sceneItemLocked: false,
      sceneItemTransform: [Object],
      sourceName: 'Note Text',
      sourceType: 'OBS_SOURCE_TYPE_INPUT'
    },

If I do the following call, it updates the source properly:

return obs.call("SetInputSettings", { inputName: "Note Text", inputSettings: { text: "Hello World" } }); 

When creating a raffle with the option "Pick a Sub", there is no announcement in the chat. There could be one too in the events column like for the classic raffle, might be useful?

The viewer count icon is not showing properly:

There seem to be an issue with the SVG URL escaping:

It seems the SVG is not encoded properly with encodeURIComponent.

For now the application is English only. As community asked, it would be interesting to provide multi language support.
I did a first implementation of this translation (FR only for now) and propose my service to fully implement FR support.
Other language would be way easier to implement after this.

Although this interface looks very promising, it lacks some features to be really usable by people with disabilities, and especially blind people using screen reader.
Here is a list of issues I found regarding this topic:

• All buttons should be labelled with what they open/do (settings for the settings button for example)
• • Within settings, there is no way to know if a box (a feature) is enabled or not: Either use a html checkbox to represent the checkbox, or, if that's not possible, you may add a "disabled" or "enabled" text alongside the feature's title, so that the information is rendered.
• • When a chat message arrives, could be nice to use a aria-live region to read it. A way to do this is using a <p role="alert"></p> and when an event occurs (such as a chat message), adds the corresponding message within this tag for about a few seconds to let the screen reader read it. after the message is read, the <P> tag can be reset to an empty content.

I'll try to see if I can create one/more PRs to enhance this. If so, they will be marked as "a11y" as it concerns accessibility.

Some streamers accept (and provoke them) copy-paste waves on their chats (for example: shaunz) which makes it unreadable.
It might improve the usability and ease the management, if some detection and clustering could be done based on a string matching.

Example of implementation idea (for a better explanation of what I mean):

There is a kind of buffer of last M messages (disappear after M messages or after a given time?).
Each new message is checked against this list.

• If it is found, a repetition counter is incremented and the disappearance mechanism is increased.
• If the repetition count is higher than N times, the clustering occurs. It could present the initiator of the copy-paste wave and the following sheep.

The clustering itself, means removal of matching messages from the 'normal chat' into a dedicated space. This space could be a clickable 'View current Copy-paste waves' button (hidden/inactive? when no wave in progress). A click on it opens a small overlay with the list of running copy paste-waves and the interesting information.

Context:

• Some users use command with text after to troll the streamer
• If the streamer has set a command block rule, all messages starting with the command text will be hidden.

Suggestion:

• An option to precise if the streamer wants to hide all messages, or only those starting and ending by the command text, would be the solution
• Technically, it should be a condition to add a regular expression anchor ($) at the end of the filtered text.
• For the implementation, with the current UI, you could simply permit to type "$" after the command text to block, in order to enable the option. (not perfect but efficient)

Edit: typo

Hello,

I don't know if the issue come from twitch or twitchat, but i've a error 400 when i try to authorize the link between both.I've already do an authorize few week ago without issue to test and cancel it until a new try

Thanks you for the tool