Tasker + ChatGPT + Elevenlabs = Super Realistic Assistant!
Вставка
- Опубліковано 5 лип 2024
- Create your own assistant with super realistic voices! Check out my sassy assistant!😆 • Sassy Assistant Workou...
Import the project in this video here: taskernet.com/shares/?user=AS...
Make sure to read the project's description on TaskerNet so that you know how to use it properly.
Download Tasker here: play.google.com/store/apps/de...
Comment on this Tasker project here: / dev_super_realistic_vo... - Наука та технологія
Create your own assistant with super realistic voices! Check out my sassy assistant!😆 ua-cam.com/users/shorts70CJOjEsTwI
Import the project in this video here: taskernet.com/shares/?user=AS35m8ne7oO4s%2BaDx%2FwlzjdFTfVMWstg1ay5AkpiNdrLoSXEZdFfw1IpXiyJCVLNW0yn&id=Project%3AElevenlabs+Voice+Synthesis
Make sure to read the project's description on TaskerNet so that you know how to use it properly.
Download Tasker here: play.google.com/store/apps/details?id=net.dinglisch.android.taskerm
Comment on this Tasker project here: www.reddit.com/r/tasker/comments/11l3f5u/dev_super_realistic_voices_in_tasker_with/
Task "Download and play sound file" doesn't work. Downloaded file isn't playable by any player..
@@bliblabl8149 Maybe the sound file you're trying to download is not valid? Can you please try opening the file in a text editor and see what it shows?
Hi works well, awesome stuff. But there seems to be an issue with long replies. Like when I ask to say a poem I'm getting the error below. Any fix ideas?
Tasker Action Error
Action: 'HTTP Request' (step 4)
Task: 'Text To Speech Elevenlabs
("detail":[ "loc".["body ",48],"'msg"."Invalid
control character at: line 2 column 47 (char
48)"type"."value_error.jsondecode""ctx"
f"msg"."Invalid control character at""doc"."(
I"text\": I"Of all the things that I have seen,
The
world, it seems, is evergreen,
With hills that rise
and valleys low,InAnd rivers that through them do
flow.In
The skies above are vast and blue,InAnd
clouds that sail like ships so true,
The birds that fly
with grace and ease,
And sing their songs among
the trees.
The sun that shines with warmth and
light,InAnd stars that twinkle in the night,InThe
moon that casts its silver glow,InAnd tides that ebb
and cease to flow,
Yes, all these things are good
and fine,
And make this world a place divine,
So
let us cherish what we have,
And always strive to
be brave.
For life is but a fleeting dream,InAnd
nothing's quite what it may seem,Inbut if we try and
do our best,InWe'l find some measure of happiness
I"
)"pos":48,"lineno".2,"colno'.47))]; (Error Code: 1)
@@joaomgcd how to change eleven lab APIs from tasker without re-importing the task?
@@bliblabl8149 me too, have you fix it?
That's honestly amazing! I'm so glad you made this wonderful app
Also, the Violent assistant was hilarious!
Haha thank you, glad you liked it! 😅
WAY TO GO! This is unreal sounding! Keep up the good videos and updates!
That's the plan! 😁👍
OMG . Those voices are amazing , you can actually hear the breathing and intonation , and they adapt it to the mood , I'm shocked !
When I saw your name and voice on it, I had to double check we are not April 1st ! x) Thanks for letting us know about it , will have some fun hehehe
Yep, they are absolutely incredible! :) Glad you're enjoying it!
What did he use to make his 9wn voice accesible?
@@brunoacosta1948 In Elevenlabs you can clone your own voice in the VoiceLab: beta.elevenlabs.io/voice-lab
Hi João, I used to be a big fan of Tasker and used it extensively, but after switching to an iPhone, I had to stop using it. However, after seeing the amazing new plugins you've added to Tasker, I'm seriously considering buying an Android device just so I can use Tasker again! Keep up the great work!
If you haven't used Tasker since João took over its development, you'll probably be blown away by how far it has come - and it was probably the best app on Android even back then.
Thank you! I actually remember you from back in the day 😁👍 Welcome back! BTW, this is not actually a plugin, just a project that you can import into Tasker...
Why would you downgrade to iPhone. Your missing out on so many things.
@@ziozvevo9751 I commented above but I had no choice, I had to get one (borrowed)
@@ziozvevo9751 missed out on a ton of apps and features by having an android too, it's like I need both to have a decent set of features and compatibility. LiDaR was a big plus for the iPhone when I bought the 13 Pro.
wow that actually sounds amazing. Let's hope to see soon enough some open-source alternative this close! good joob joao
Thanks a lot! :)
Bro, it's just a few bux, and you won't be disappointed. Plugins aren't much either, and allow you so many more options
Thanks so much for this project! As always, great work!!!
No problem! :) Glad you like it!
absolute straight forward tutorial, not too much info, not too less. was able to completely replicate. amazing, thank you!
there are some minor bugs, like that it searches for your very voice model on the test voice widget, or it doesn't update my shortcut after voice change and i need to recreate the widget, but its just minor not important bugs.
continue com o excelente trabalho, sou fã do Tasker a anos!
:) Obrigado!
I'm so grateful , Soo grateful
Glad you like it! :)
I am having an absolute blast with this setup. Major props for integrating these api's onto your app platform. I forsee big things ahead for you sir.
Side note: using elevenlabs is superimpressive, but quite costly. And yeah i wish i could run that as my tts.. its just way too expensive, even if licensed I can't see the value.
Some of the other tts engines are fairly decent if manually upgraded (samsung especically) and its better than google key. You'll need to keep the default on 6th action but it will still call for google api key, not sure how to disable that yet.
But you sir, are a God amongst men.
Thank you for what you're doing.
Next best thing? Integrate a face with animations
Thank you :) Glad you like it!
thats insane, good job mate.
Thanks! :) Glad you like it!
you are just great. you have helped me extremely with my realization
Happy to hear that!
I'm also a big fan of tasker, I can't describe how powerful and user-friendly this app is
😁👍
Omg 😱, can't believe you added your own voice. Htf did you do that! 👌👌👏 awesome stuff!
Thank you! :) You can do it on their website: beta.elevenlabs.io/voice-lab You upload a few good quality samples and you're good to go!
@@joaomgcd did you pay 11labs to make your own ai voice?
@@GiathiKiarie Yes, I did :)
Demais cara! Sensacional, estou importando agora mesmo, parabéns pelo conteúdo, vc é fora da curva
😁👍
I'm waiting for this bro. I know you would bring it to us. Whenever Eleven labs announced those voices, the first thing that came to my mind is Tasker :)
It became true now..
Thanks bro @joão Dias. Keep going. All the best
Thank you very much! :) Glad to help!
Amazing stuff! Thanks!
Glad you liked it!
@@joaomgcd you're a legend man! I wish you all the best.
@@NikolayMegdanov 😁👍
OMG I love your work!! Thank you for sharing it with us.
I wonder how can you train your custom voice on ElevenLabs so natural. Any tips on your training scripts (contents, vocabulary, length,etc)? Because mine do not sound that alike...
Also any other suggestions to connect ChatGPT with Elevenlabs if not use it on Tasker? (I'm currently trying with Python, would love to know if any better ways or platforms) Thank you so much!
Thank you! :) To train the voice I simply uploaded some random samples from my previous videos 😅 I didn't take into account anything special, just uploaded something like 20 or so clips. I did notice that it got a bit better every time I uploaded some new clips, so maybe try that.
To connect ChatGPT and Elevenlabs you just need to use their APIs like I did with Tasker :) Check out the Tasker tasks or the API documentation to learn how to do that.
Hope this helps!
You Rock bruh!! 🔥
Thank you! 😆
Hello João, I just installed Tasker and your awesome project, thank you! As I'm completely new to Tasker, I have some difficulties accessing the settings for the Elevenlabs Task. How can I have the settings easily accessible from the homescreen, as you seem to have? One more question please: how can I interrupt/stop a running voice output of Tasker/ChatGPT/Elevenlabs? Thanks so much.
To add to the home screen you have to long-press your home screen > Add Widget/Shortcut > Add Tasker "Task Shortcut" widget
About interrupting the voice output, you either use the "Shut Up" action for ChatGPT or "Media Stop" for Elevenlabs :)
Hope this helps!
Dang, that's just awesome.
😁
omg this is crazy good ! - thank you
Glad you like it!
awesome man 👍
Thanks ✌
this is so cool, I watched a YT video about some new Android Devices building chatGPT into their assistants and then stumbled on this video - this is so much more customizable too! these tasks are a lot more complex than my usual tasks within tasker, is there a way to have the voice responses also displayed as text on the screen?
Yeah! :) You could add a dialog or a toast or even a notification on the screen with the text :)
Someone is in a bad mood here..
That "WOW" was hilarious tough ahah
Nothing to say, Open Ai and Eleven Labs have done a great job.
Yeah, it truly is next level!
This is amazing.
Thank you! :)
@@joaomgcd No, thank YOU.
Wow, that would be so cool to use for a system wide tts, I'm blind and just got tasker the other day. Still trying to figure out how to use it. Lol
Yeah, that would indeed rock! 😁 It would be expensive though, to use it everywhere!
Hey João! I'm enjoying using Tasker to run my chatgpt voice assistant. One thing I want to ask you is how I can stop the assistant while it is giving me a lengthy response. I haven't been able to figure out how to stop a response before it is fully completed. Any tips on that would be great!
Hi. For that you need to use the "Music Stop" action :)
Elevenlabs is so expensive. $5 isnt much, but I used up half my monthly allowance in 2 hours. The next option up is $22/mo and you actually get less characters per dollar...which you'd think it would be the other way around. Its $5 for 30k characters and $22 for 100k. 🤔 anyways.. I had a lot of fun playing with this. Sadly not something worth paying 22/mo or more for. Hopefully we'll get a cheaper alternative to eleven labs soon.
I also spent more than half of my credits in one night 😂
First of all thanks for showing us this. My mind if blown with the possibilities. I did run into an issue though..Any idea why sometimes the reponses are not spoken? They come up as written. Also would it be possible to give this a wake word like google assistant?
Do you get a notification on your notification list what that happens, by any chance?
@@joaomgcd same issue happening with me.. I think if it s a long response it breaks the read out. For example ask it to tell you a story about the ancients Mayans. It'll break it every time
@@DaveSmart28 I think that's a limitation of the Elevenlabs API. It doesn't allow you to generate text to speech with a very large numbers of characters unfortunately.
@@joaomgcd that's disappointing.. :(
"How about food that- that's usually a good option for lunch" made me fall out of my seat
😅
hi, i have a question about delay: when I launch tasker and I ask a question, it takes like really a long time before it replies with an answer. In your video, it's like almost instant. Do you think I can adjust something to make it work more quickly?
In my video I cut it a bit short to make the video less boring. In reality it takes a few seconds for me to get the response as well. Unfortunately that's because voice generation in Elevenlabs is a bit slow still. Maybe they'll improve it in the future!
I saw you reply to every comment, really thank you I have a question too; It takes too long to get a reply after I talk and send it, how can I solve this? if i can't figure it out, is it because chatgpt is read after the answer is finished? if so, can't it read at the same time as the answer? For example, I can use some google chrome plugins with kiwi browser, one of them is the chatgpt plugin, this plugin starts reading without waiting for the answer to finish. Is it possible to integrate it or use such a feature? (note: even without natural voiceover the answers are late, probably not the voiceover)
Unfortunately that's not possible with this API, sorry! You really have to get the full response before sending it to Elevenlabs....
awesome
@João Dias Is Elevenlabs free of charge or can I use my phone's orginal voices? Thanks!
You can use your phone's voices if you want. This is just a way to use Elevenlabs instead if you want :)
Being able to do this in 2023, in a production phone, by an individual, is MAGIC!!!
Cool! 😁Glad you like it!
Hello talking in money terms does this gpt chat ai and the other one costs? Or they can be used for free
Yes, they have costs. You can find out more about their costs in their respective websites :)
Hi. This is my first day using ChatGPT and ElevenLabs and Tasker. Could you please give me step-by-step instructions on how to set this up as my default assistant, and how you program her like you showed above?
Thank you.
Hi! :) Do this:
- Open Tasker
- In the main Tasker screen go Menu > More > Android Settings > Assistant Settings and set Tasker as your assistant
- Go back to Tasker and go to the "Profiles" tab at the top
- Add a new profile with the Event > Assistance Request condition
- Go back to accept
- Select "Voice Chat ChatGPT Elevenlabs" as the task
- Back out of Tasker with the back button all the way to save
It should work now! :) Hope this helps!
@@joaomgcd This is great, but it doesn't seem to be opening any apps or doing assistanty things. It *DOES* however pop open the voice interaction.
Can I also get it to function as my assistant? I went through the steps and got some commands enabled, following the instructions on your share.
@@siigari oh, it won't actually do anything on your phone. It's just a chat bot, sorry 😅
Hi João! nice video! so funny! i know it migh sound dumb haha but how i put the icons on the screen like that? so i can use manually as a common app?
the tasks in this case
Long-click your home screen > Add Widget > Tasker Shortcut :) Glad you like it!
How do you get responses so fast? Sometimes it takes up to 30 seconds to get an answer
I had to cut out the wait times for the video or else it would be pretty boring to watch 😅 I would normally get something 5-10 seconds wait times though.
after one or two prompts that i had received elevenlabs successfully, then sometimes the next response will not be voiced out and it seems like it was feeding back to google voice input as I see in the text, then google voice said "Didn't catch that. Try speaking again". How to solve this ?
Do you get an error in your notifications when that happens?
hey I appreciate these videos.I want to do everything on the this video. How would I go about about setting up chat GPT with Tasker. Can you direct me to a video where you of walked people through this. I did read the article but it's so much easier for me to follow a video?
Hi Kevin. You just have to follow the simple instructions in the link in the video description :) If you really want a video you can check this one (not made by me): ua-cam.com/video/Q0_7RSOsA2s/v-deo.html
this is insanely cool - I sometimes get an error message if there is an invalid character in the response from Elevenlabs, and it just doesn't say it afterwards. It does display the response text though however - anyway around this?
Thank you! :) Can you give me an example of when that happens? Thanks!
@@joaomgcd Action: HTTP Request Step 4
Task: 'Text to speech ElevenLabs"
{"detail":[{loc:"["body",134],"msg":Invalid Control character at: line 2 column 133 (char 134)","type":"value_error.jsondecode","ctx":{"msg":"Invalid control character at","doc":{
\"text":\ ..................chatGPts response.....
(error code: 1)
Usually occurs when it attempts to say numbers, or letters, page breaks, table contents etc.
@@reci2682 Thank you! Could you please create a small test task that simply calls the Elevenlabs Text To Speech task and generates that error? Thanks again!
I'm also running into these issues. You can test this yourself by asking it: 'Recommend me three gifts for my mother's birthday'. I get the error every time. I've posted the log on Reddit in the ChatGPT topic.
Is there a way to trigger this with a trigger word, just said out loud (like google assistant and alexa etc) or do you always have to press the shortcut to start chatting?
Hi! Check here: ua-cam.com/video/LpX1hNMxVsI/v-deo.html
Hey I got mine working great but seems to be very slow compared to your video. Are you speeding it up at all? Any suggestions how to make it run a little quicker?
Yeah, I sped it up for the video so it's less boring to look at, sorry about that 😅
@@joaomgcd All good, still having great fun with it. Thanks for sharing!
Man I haven't been excited about technology for a while until I got to play around with gptchat. I've also been outside the Android ecosystem because of a fatal flaw that has no known fix ( trust me ) on Google Pixels, they don't receive phone calls! What good is a phone that can't function as a phone? I'll dig up my old pixel just to play around with Tasker again, I really did miss all the cool things you can do.
Hi João, my voice widget wouldn't loop and gives me just one answer. Do you know how I can fix it?
I can run this from within tasker but I dont know how to create shortcuts so I can change voice etc. The way you went back to your phone homescreen and selected voice I dont know how to do that from within tasker. Its great I just wish I could change voice or add shortcuts the way you have it set up
Hi. You have to long-click your home screen > Add Widget/Shortcut > Tasker Shortcut > Choose appropriate tasks.
Hope this helps!
@@joaomgcd I will try that, thanks!
I am not a coder, but would it be possible to add an extension to this projetct like web chat gpt to enable internet search as well?
Not very easily, but in theory you could use the HTTP Request action to get some info from the web and then feed that into ChatGPT :)
Do you have to type everything out or can you speak to it like you did in the other video?
It shows me talking to it in this very video 😅
@@joaomgcd No you did not or put the timestamp for when you did. You just typed or copied text into the box and then the voice assistant responded, but you did not say anything out loud. You must be thinking of your previous video.
@@sadbravesfan He muted the parts where he was speaking. He was not typing in his requests but speaking it out loud.
@@sadbravesfan Oh, sorry, I did speak to the assistant in this video. It's just that since it's a screen recording I wasn't able to record the voice on the video at the same time as the popup in Tasker captured my voice. 😅You can't hear it but I did speak to it in this video.
How you make this homepage folder? Project file does not have this....
You create multiple "Tasker Shortcut" widgets by long pressing your home screen and adding them. Then you drag them on top of each other to create a folder.
What can be done about delays in response? It seems like the answers takes ages to return a TTS response.
Not much we can do unfortunately. Just wait until the Elevenlabs service is better optimized and is faster.
@@joaomgcd in it's current stage, it's unusable due to how long it takes between request and response. In the future this has the potential to be extremely powerful though. Thank you for sharing this with everyone.
How can I set any of the speech synthesis as the default tasker voice engine?
Just use the provided task to say your text instead of the "Say" action :)
Cool!
Say, can I ask how did you grouped those tasker tasks shortcuts together on the home screen ?
Long-click home screen shortcuts > drag them on top of each other :)
Thanks for the reply, for some reason it works for apps but doesn't work for me for tasker shortcuts icons
@@guyossoffer1856 Are you sure you added the widget called "Task Shortcut" and not the one called "Task"?
With Xiaomi miui 13, it seems as if I have two options, one icon with a timer and the other one it's probably a task widget . and there are no more options. If so I'm doing something wrong
@@guyossoffer1856 Seems like your launcher doesn't support shortcuts then. Can you please try a different launcher? Thanks
Very useful video!
Since chatgpt has now given the ability to talk to the bot directly in the app, could we create a profile linked to a specific button (on the phone or headset) that opens the app and clicks on the voice chat?
This way we would avoid using the chatgpt API and save in the app all the voice chats we have done.
I hope I have explained myself well and I know you will be able to do it! 😜
Ok, that can be done! Go into Tasker > Create Task > Add Action "Launch App" > in the app selector long click the ChatGPT app > Select the "AssistantActivity" option. Then enalbe the "Exclude From Recent Apps" and "Always Start New Copy" options and accept. If you run that, it'll open ChatGPT in voice chat directly :)
@@joaomgcd Okay thank you very much!
I couldn't find the voice assistant function.
I tested this task a bit, and I noticed that a problem arises: the assistant continues the voice chat to the bitter end at least until I manually close it.
Could we add a function, like in the video, that by saying a word, for example stop, he makes it close?
Also, how can I make it start this task by associating any key on the phone (e.g. volume up) or a key on the bluetooth headset (e.g. volume up, or the 2 taps) but with tasker closed and the phone locked?
@@gaetanopalacino Unfortunately there's no way to control that, since this is just using the ChatGPT's voice chat functionality and I can't change how the app works, sorry!
First, thanks to this amazing fun project !!
Where did I go wrong ? I installed everything like told. Chat GPT itself works so far. My problem : I can ask one question and then get the answer, but then the Google Voice Assistant does not pop up anymore. Or it sometimes pops up when it answers, and instant starts Google Voice Assistant so I can't even read the answer....
Hope there is a solution to my problem.
Thank you! :) Are you doing that with the Voice Assistant task in the project? :)
Is there a way to enable the assistant to use the internet? Like tell me the weather in my area
Yes, you can use Tasker's HTTP Request action to get whatever info you want from the internet and then for example send it to ChatGPT to make a summary of it :)
Hey Joao - This is awesome, and I believe I've got it working. However, I can chat with it (eleven labs), but whenever I ask it a question "How do I make xyz?" or "list me xyz" it throws up an action error notification, "invalid control character, line 2" etc, and does not respond. Any clue how fix this?
hhmm, interesting. Do you think that lists will include some kind of character that's creates that issue? Do you know what the character is?
@@joaomgcd thanks for replying! Well, it's not just lists, it's most things where I ask for an explanation. It shows the text in the dialogue box as if the bot is speaking but no sound and I get the error. Is this something you have not experienced on you side?
@@bccbear_ Maybe I just didn't get the correct responses for that to happen😅 Can you maybe use the "Input Text To Elevenlabs Speech" task to manually reproduce that error? If you're able to, can you please send me the text you used so I can try myself? Thanks!
I got this working, but missed the description for how the AI should behave. What do I need to edit it to modify it's behavior?
NVM. I was editing the print and not the action, just had to create a new widget and bind it to the assistant personality.
Im using an s10e and was able to find this task to double tap Bixby button as well, so I can jump straight to scarlet johnasson via Google assistant via chat gpt lol
Nice! 😁Glad you got it working!
@@joaomgcdsadly it immediately became obsolete. The services are all paid, but chatgpt has its own voice now which is free. Tasker still works if I disable elevenlabs, but uses Google text to speech instead of chatgpt's. I wonder if there's any way to use tasker so I can map chatgpt's own ai
Fantástico João Parabéns. No entanto estou com alguma dificuldade em por o assistente a falar pT-PT. Estou a usar o "Say WaveNet", já alterei a "Voice" para "pt-PT-Wavenet-A"... mas nada acontece. Está-me a falhar alguma coisa? De resto fantástico!
Olá! Obrigado! :) O que queres dizer com "nada acontece" exatamente? A tarefa não corre sequer? Podes tentar carregar no Play dentro da tarefa (dentro do Tasker) a ver o que acontece? Obrigado!
@@joaomgcd desculpa. Não fui claro. A tarefa corre mas fala em inglês. Não o consigo é por a falar em português. Se eu pedir para ele me dar a resposta em português ele dá... Mas com pronúncia inglesa, o que não deixa de ter a sua piada. Parece que estou no Algarve a falar com um estrangeiro😁
@@joaomgcd No ponto 6 "Say WaveNet", no item "Voice", alterei para "pt-PT-Wavenet-A", mas ele continua a falar em inglês!
@@lemi00 Hehe, também já experimentei isso e é realmente engraçado 😅 Mas será que o que está a correr é a ação "Say" e não a "Say WaveNet"? Já inseriste a API key para usar com a Say WaveNet? Se correres a tarefa "Voice Chat" dentro do Tasker vês algum erro?
@@joaomgcd Sim, já coloquei a API key. Quando corro a partir do tasker (carregando no play) não dá erro algum. Agora, por milagre (após alguns resets) já usou a voz portuguesa mas responde em inglês. Se eu pedir para responder em português ela responde na perfeição. Obrigado pela ajuda. Será que se eu puser na "Personalidade" para ela responder em português ela o faz por defeito?! Vou continuar a explorar. Não te chateio mais. Mais uma vez, obrigado pelo fantástico trabalho.
Hey whenever I use the eleven voice chat in tasker it seems to reset the personality I have set in chatgpt. Is there a way to fix this? Sorry I am new to tasker.
Hi! That does sound weird! Are you using the shortcuts on your home screen for that?
Elevenlabs is offering multiple languages now, but I have not figured out yet how to switch to German. Is there away to modify the http request so that I get a response in, e.g., German?
Ok, I've modified the project :) If you re-import it, it should allow you to select multiple language mode now!
Hi Joao, wich Version of Chatgpt i use in voice Assistent, i am a plus suscriber.
It's 3.5, but you can change it in the "System >> Send Chat" task in the "HTTP Request" action :)
@@joaomgcd thanks man😍
한국인을 위한 설명: 유튜버의 바로 전 영상을 보고 ChatGPT 프로젝트부터 import 한 다음 이 영상의 음성생성 프로젝트를 import 해야 작동함. 처음에 뜨는 대화상자에 오픈AI API키를 입력함. 그리고 OpenAI api 빌링 웹페이지에 들어가, 결제수단 등록해서 사용량(quota)을 얻어야함. 수단 등록 안하면 api 쓸수가 없음. 다시 음성생성 프로젝트로 와서 Voice Chat 으로 시작하는 태스크 실행하면 잘 됨
Thank you for the explanation! :) The instructions are in the TaskerNet project as well!
Hello, any tutorial on how to do this? been wanting to have chat gpt as my assistant
Hi. The project in the video description describes how to get this up and running :)
Can you use ur own made voices or r those samples like Rachel and adam the only options allowed
You can use your own made voice, but I think that's a paid feature in the API.
@@joaomgcd understood
Thanks heaps for making this tutorial. When I hit the voice chat button it brings up Google voice recognition, which seems to cut out for no reason. Any thoughts or solutions? It seems that Google speech synthesis and recognition is my only choice for voice input
It seems that you have to keep talking or it will think that you already concluded your command. Unfortunately there's no way around that I don't think, sorry!
I think my issue is a little different, when you successfully speak it will make a little success sound, like with a rising tone, but when I finish speaking it makes an error sound, like with a falling tone, and doesn't tell me why
@@Mp3Robbo hhmm, that's weird. Did you maybe block mic access for the Google app or something like that?
@@joaomgcd well it spells out what I say fine, like the entire sentence of what I'm speaking underneath the google microphone icon, so it must have access to my mic, but then it gives the error sound. Interestingly it seems that the message still sends to gpt and I get the response, but then tasker complains that it can't find the audio that was received, I struggled to screenshot the error message but I know that tasker is trying to store little mp3s on my phone with each response
@@Mp3Robbo Hhm, that's a different issue then. To do the Elevenlabs TTS Tasker writes the sound file to your device and then plays it. Can you try running the Elevenlabs test task and see if that works?
How can I use the Turkish voiceover language option?
When you import this project you have to run the setup task and there select that you want to use multiple languages. Then simply send something in Turkish and it should work (if Elevenlabs supports Turkish)
Absolutely amazing.
How did you do your own voice though?
beta.elevenlabs.io/ has an option to do so! :)
@@joaomgcd you know thats not free how can we bypass those payments and use our voices as our own ai voice
@@GiathiKiarie you can't
@@milazapallo8995 yes because 11labs website asks me to pay some amount of money to set my aivoice,how did you create yours without paying help please
@@GiathiKiarie it's not possible to do that without paying (yet)
I keep getting invalid api key errors, its also not allowing me to set it to GPT4 however my subscription is active
Did you correctly set both the Elevenlabs and OpenAI API key when prompted?
Is there a way to use another voice recognition system other than Google's. It's frustrating to be interrupted just because I stop speaking for 1sec. It would be great to use a voice recognition system that records what you say, like in a whatsapp audio...😕
You could use Tasker to create a scene with a button to start/stop recording your voice and then send the recording to OpenAI's Whisper API to get the text: www.reddit.com/r/tasker/comments/11zsuxw/howto_transcribe_text_with_openais_whisper/
Hope this helps!
Thank you for your answer!
Though, it would be great to know how to do that.. I've tried for hours without success..😭
I dont want to use with chatgpt, just to change the voice used in other tasks/projects. Is it possible? If so, how do i do that? Ty!
Yes, just use the Elevenlabs project and then use the "Text To Speech Elevenlabs" task in any other task (with the Perform Task action)
Is there any way to get this to work on the galaxy watch?
Not that I know of, sorry!
Como puedo hacer que hable con la versión en español de esas voces?
hi. I just updated the project :) Import it again and it'll ask you if you want to use multilingual mode. Hope this helps!
Hi João!
I need to know how use this but with a button to start/stop recording my voice. If you could please help me out with this? I can pay you in exchange. This is for a project I'm working in to learn languages!
Ok, I can help :) Please send me an email so I can better help you out! Thanks!
everything works, BUT i do not have prepaid chagGPT, i chosed 3,5 version but is asked paid version anyway, and now only bliks the screen, but i won"t get answer.. any help or explain?CAN IT WORK WITHOUT PREPAID PLAN chatGPT?
Hi.Tasker simply uses the ChatGPT API for this, so if you want to use it, you need to pay whatever is necessary for the API to work :)
@@joaomgcd anyway I LOVE YOUR WORK. in another video you gave me, what I was looking for a few weeks. Keep working. every your post works like a charm. giving subscribe. and I think i will pay one day chat gpt. I would have some question, but it is not public.... 😉
I like it but still needs work, like being able to add it to the say task for easier adjusting options, and the tasks are slower to run
That's because this is using the Elevenlabs API to generate the voices :) It's not actually a new action inside Tasker itself. The API is just slower to generate the voices. Hopefully they can be quicker in the future!
Wow, awesome project! Is there a way to set this as the default assistant? I noticed you can set Tasker as your default assistant but wasn't sure if there was a way to auto run this voice action
Update: Figured it out! You can make a Profile for Event UI/Assistant and assign this task to it works like a charm!
Yep, that's it! :) Glad you got it!
Could you please share how you did this? This is my first day using any of this stuff! I have the ChatGPT and Tasker scripts installed.
@@siigari Did you figure it out in the end? Let me know if you have any guidance as I just saw this video and am going straight into setting this up, as I wanted sth like this yesterday!
Yes please, how is this done? I found how to start Tasker Tasks from Google Assistant, but not how to *replace* Google Assistant with ChatGPT. I can't even locate Event UI/Assistant in Tasker. Does the device need to be rooted?
Ref
How can I make google assistant use elevenlabs? I have all the chatgpt projects and the elevenlabs project, but I can't put them together. The task chatgpt elevenlabs works, but when I say 'hey google' it's the bad tasker voice. I've tried disabling 'say' in 'hey google chatgpt' in the task 'chatgpt voiceloop' and put perform task - text to speech elevenlabs with parameter 1 set as %response, but when it just doesnt work. Please, what am I doing wrong?
You have to replace the "Say" action in the task with a "Perform Task" action that calls the "Text To Speech Elevenlabs" task and set Parameter 1 to the text you want to say out loud :) Hope this helps!
@joaomgcd Yep, did that. I guess my questions would be: 1)is it right to change par1 as %response if I want the hey google to answer using elevenlabs voice? and 2)is it the 'ChatGPT VoiceLoop'task under 'Hey Google ChatGPT' I want to change? Just to be clear - Elevenlabs ChatGPT works just fine and Hey google assistant works fine as well. What I'm failing at is saying hey google, saying something and getting answer said by elevenlabs (even if I don't have tasker opened.) Again - sorry, really new to this thing.
@joaomgcd I'm really good at following your videos and instructions, it works, I'm just totally horrendous in programming it by myself 😅
@@MartyOfEarthOk, so, you imported this, right? taskernet.com/shares/?user=AS35m8ne7oO4s%2BaDx%2FwlzjdFTfVMWstg1ay5AkpiNdrLoSXEZdFfw1IpXiyJCVLNW0yn&id=Project%3AHey+Google+ChatGPT#
Go into that project and edit the ChatGPT Voice Loop task.
There you'll find a "Say" action, right? In the same place as that action add a "Perform Task" action to call the "Text To Speech Elevenlabs" and set parameter 1 to %response. Then delete the "Say" action.
Back out of Tasker all the way to save and it should now use Elevenlabs instead of Tasker's worst voice :)
@joaomgcd OK, not really sure what I did wrong, but it works now 😅 Thank you so much, you're amazing! Not only for creating all of this, but also for keeping up great customer support 😀 Thanks!
How can I use ElevenLabs TTS engine in my Android Studio project for TTS service in my app
Check out their API. You can use it in any app that way.
The response time for the voice assistant is really long, is there a way to make it faster?
Unfortunately not, sorry! The slowness comes from the API, not Tasker.
@@joaomgcd No problem. If I upgrade elevenlabs to starter tier, do I have to change anything in tasker to recognize it? Same with chatgpt plus. I have no idea about this stuff so I appreciate the help
@@rumak9818 Nope, all the API access stays the same :)
Will you release it on Google Play Store?
Tasker is already on Google Play Store :) You can simply import this and use it in Tasker right away! :)
This is good, because google api has started giving me fits.
Yeah, always good to have an alternative.
Sometimes talk and sometimes writte, someone know how to fix it? Appears error 1 and after that: couldn't get the result
You figure this out?
Hey man, I was wondering if you could help me out with something... I have created a voice assistant and want to use a voice I have created in ElevenLabs but I can't figure out how to do this. I'm using python and Visual Studio Code... Do you know how I could add the voice?
Use their API :) The Tasker project in the video description shows you how to call it so you can check there if you want.
@@joaomgcd It does but only via tasker, I want to import my voices into a python environment using Visual Studio Code
@@Haks117 You have to use the same API via Python :)
@@joaomgcd I have used my api and imported the elevenlabslib package yet still doesn’t work
@@Haks117 Sorry but I really don't have the time to guide you through Python programming... 😅hope you understand!
Maybe I got this wrong, but I think the English model uses Eleven monolingual model, then I tried to change it to the Turbo V2 model manually which I think is faster but it doesn't work. Any ideas?
Do you get an error when doing that?
@@joaomgcd Yes, it seems to never find a sound file to play.
@@BrunoHenrique-ov1le What happens if you run the Preview task manually inside Tasker?
@@joaomgcd Well, now it seems to be working. Don't know what was going on. It just wasn't working with the turbo model then I would change it back to monolingual and it would work. Sometimes I just get some errors, like ChatGPT connection reset and things like that.
Took me a while to test it as I was not using elevenlabs anymore, it is too expensive.
Just tested and it's AWESOME! Will this work on AutoWear?
As of now you wouldn't be able to make AutoWear speak out the voice, but I'll try adding that to AutoWear later :)
@@joaomgcd this would be great, auto apps are the best purchase I ever made
Nice work! Thanks for your video, now I can use it too! But why are you Rachel so emotional, I just want that type of voice, can you teach me how to set? My Rachel voice is totally different with yours.
Glad you like it! :) To get voices to act like that, in the beta.elevenlabs.io/speech-synthesis website configure Rachel's voice and set it's stability to 0% :) Let me know if that helps!
@@joaomgcd WOW! It's wayyyyy better, thanks alot body!!! But the way, where is the chat history file of the chatGPT Project in your video?
@@zenry6124 Great, glad it's better! About your question, what do you mean by "chat history file"? Thanks!
@@joaomgcd "Oh, it's my fault. If there's no history file. It's because I used a different project called "ChatGpt for Android" in Tasker. This project has a "ChatGPT history" file, which contains a TXT file with all the chat content."
@@zenry6124 Oh, I see :)
I told voice gpt to tell me a scary story and it errors out. How come
What error do you get?
Can you use your chatgpt plus login?
What's that exactly?
In the Voice Chat ChatGPT Elevenlabs task, I got a error that "Task doesn't exist:System >> Clear Chat". What should I do to fix this?
Did you install the ChatGPT project as mentioned in the TaskerNet instructions for the Elevenlabs project? :)
@@joaomgcd I will find that. Thanks! Could you make the tutorial include all the process needs?
@@자비스 All the process is explained in the TaskerNet descriptions of the projects :) Did you check those?
@@joaomgcd I missed the description, And I've found out it is needed to set up OpenAI API billing for usage limit. Im making progress. Thank you so much!!!
@@자비스 Awesome :) Glad you got it!
João alguma possibilidade de usar esses assistentes em português brasileiro?
Infelizmente o Elevenlabs só suporta inglês. Mas o Say Wavenet no Tasker suporta Português, portanto se usares isso dá para ser em Português! :)
Elevenlabs está oferecendo vários idiomas agora
@@Lua-di4vf Sim, se importares outra vez o projeto ele vai dar a opção de usar com outras línguas :)
Can I use this to read my whatsapp messages out loud on my google home mini speaker?
You could use the AutoCast Speak action for that.
Hi there. This video is cut right? Bevause i wait nearly a minute to get my response
Yeah, the video is cut, but it usually doesn't take a minute for me, more like 5-10 seconds or so.
@@joaomgcd yeah I've noticed it varies depending on how much you put into the prompt and personality. Also device (mine is low end)
Hey i just wanted to also thank you. This stuff has been amazing, useful and very entertaining. It has also sparked an interest in seeing what else i can do with Tasker! Great job, João 😁👍
@@johnnywoodstock Thank you very much! :)
I just set up this task. I get a ChatGPT Error: You exceeded your current quota... Can you help?
I've got the same error. Any fixes?
If you want to use the ChatGPT API you need to pay OpenAI :) I don't control the quota myself.
If you want to use the ChatGPT API you need to pay OpenAI :) I don't control the quota myself.
I appreciate your hard work as always joao, but their TTS isn't worth that pricing structure given how much some people use assistant daily. The ai space is so fast and fluid all of these companies are just charging wild amounts because they know they could be gone in 2 months. Sad but just the way this market is.
amazing video!!
I have a problem sometimes the IA response me writing and sometimes talking? Someone know how to fix it?
Do you get any errors in your notification list when you do that?
@@joaomgcd Yes, error 1 appears and after that: couldn't get the result
@@Blue9000rpm Does it have any more details in the error?
@@joaomgcd no, the only thing I do is import the two Projects with their corresponding APIs and that's it, did I forget something? I really appreciate you answering me( my English is not my language and is the first time I use tasker, after watch the video 😂)
@@Blue9000rpm Do you see an error in your notifications perhaps? Are you running that from inside Tasker or from a home screen shortcut?
I can´t pass the api key screen, I have set up even a paid version of the Openlabs and still saying Invalid Api Key, any ideas ?
What does the error say exactly?
@@joaomgcd Error: 1 detail " status ": invalid_api_key","message":"invalid Api Key : xxxxxxxxx I have tried deleting the api key and creating new ones and nothing
@@mercuryin1 Sorry, I don't think I can help :( That's just an error that the API returns. Are you sure you're using both keys in their correct places?
@@joaomgcd I am using the api key in just one place where it ask you as soon as you import the script and click play, where is the second place to write the api key ? This might be the reason because I have tried the api with your script but from a different chinese youtuber and it works perfectly. Thanks for your help
@@mercuryin1 You need an API key for the ChatGPT API and another one for the Elevenlabs API :)
The HER film becoming a reality.... :V