Видео

thank you legend 🚀

thank you my man 💪 To do this dynamically you could create a bunch of reasoning variables (e.g., reasoning_1, reasoning_2, reasoning_X) and then write a function that parses the reasoning steps out of the chatGPT completion API response, then iterates through each reasoning step and assigns it to the variables. This does mean you'll probably need to create a lot of dummy reasoning variables, but this method should work. Hope this helps legend 🚀

Thanks legend 💪 Will do, will make some more vids around coding or building an app. And will keep things slow and simple 🚀🚀

Thanks my man! I didn't use the Zendesk Make module to keep the build dynamic incase there were non-Zendesk users following along

thank you my man 🚀🚀

No worries my man, I'll be putting out more beginner coding content in the coming vids 💪

hello legend! Nice idea, I'll look into it :) The main hurdle is that end users can end the chat at any moment, so we won't be able to "trigger" the end of the chat. A workaround is to do a weekly automation to get all of the transcripts (e.g., for the last 7 days), iterate through them one by one and classify/ review them with AI, then push the final report to something like airtable. Thanks for the ping 💪

@@BartSlodyczka Thanks for your consideration. We would find it beneficial to have such reporting on a daily basis as well, to make sure that there are no spikes in critical categories, such as login issues, payment issues, etc.

Build on this example but in different files and talk about context, updating, how to include other libraries, push to fit etc

@@telegraphavenuetv done and done, have noted this down for future vidzzz

Hello legend, you could use Voiceflow's free tier or you could code up a basic html/javascript widget (using chatGPT) that just serves a super basic convo. If I was to choose I'd say Voiceflow free tier because their widget makes the customer experience really nice. Hope this helps 💪

Hi Bart, I'm not 100% sure if I get it right - but to me it seems Voiceflow only offers a 14 day free trial period. At least it says so to me - so this doesnt seem to be an option unfortunately :(

It's 100% free :) I use the free version for my demo account. The free account does have limitations like only 2x assistants and limited AI tokens, but if you just want a simple web widget front end and nothing else, then the free tier is perfect 💪 the 2 weeks is the "pro" plan they give you for 2 weeks, after which it goes to "free"

@@BartSlodyczka ohhhhhh okay thanks, so it was misleading me to trick me into extending the pro version. Sounds great! Do you know if these ai tokens are counted for all requests sent to replit to the custom gpt backend, or of this totally free from VF-side no matter how much my customers chat with the chat ui?

haha! so the ai tokens are only counted for ai stuff you use within voiceflow. If you have a basic flow like in the vid, where the convo is sent to replit/openai, then you use the ai tokens from your openai account (and none from voiceflow). I don't use any ai blocks in my voiceflow canvas and never go over :)

Awesome! Glad this was helpful 💪💪

Thank you legend 💪

I would add a "virtual" delay by editing the global message delay in voiceflow settings. Otherwise I'd use the API example in the video

Done

This is the best coding assistant considering it is imbedded into the server platform! Good luck with your builds, legend 💪

hey legend, yes you will need to deploy your replit to make it always on. In Replit, there is an option in the top bar on the right hand side that says "Deploy". This will then make your replit live and always on. This also changes the URL so the only thing you need to do is edit the URL inb Voiceflow to this new URL. Hope this makes sense my man! 💪

@@BartSlodyczka Yeah, thanks bro for the tutorial and the response. What do you recommend for the deployment options, is autoscale grate for this kind of projects?

@@santiagoandresdoria3365 yeah that’s a good setting. If it turns out you need more (eg you reach your limit) then you can upgrade. For now autoscale is 👍

Perfect, will speak to Josna tomorrow

Send me an email: bart@supportlaunchpad.com

@@BartSlodyczka please check it out, thank you.

@@kevinsedwards interesting, I’ve never needed to add a delay after a text block, and this might be a limitation on Voiceflow’s end. What is your workflow that you need to delay after text is sent? What is your next block after the text block?

​@@BartSlodyczka I built a basic customer service sales bot that just answers questions. all I want to do is slow down the messages so it doesn't jump to the next block after finishing it's loop, letting them have time to read the message. simple stuff not sure why it's so hard to do.. thanks

@@kevinsedwards Try this "dummy step" ruclips.net/video/Pp3WkauRrQ8/видео.htmlsi=ZNtbeS2wzp6ZuTLM&t=201 Not sure how Voiceflow backend works so not 100% if this will work for you, but if you add this dummy step after your text step, and set it to e.g., 3 seconds, you should get a visual delay in your chatbot. If this doesn't work, might need to get creating with buttons and sending a "continue" button or something else. Good luck!

@@BartSlodyczka thanks man will give it a try later. the message came up as "try this dummy" lol

@@kevinsedwards hahaha that's wild

Thanks!

So we just need to copy and paste the function in voiceflow right?

@@mardini777 Yes that's right! In Voiceflow, create a new function. Add the github code into the function (you might need to delete the 1st and last line of my code - you'll see what I mean). Then on the canvas just insert a function block and select the function you created. Then map the variables that go in and out of the function. :)

hey legend, send me a loom of your build and I'll take a look (bart@supportlaunchpad.com) Also explain why you want to use Zapier because maybe it can all be done in Voiceflow I.e., what is the overall workflow 💪

@@BartSlodyczka many thanks bro ! I send you all that

hey legend, just noting that I got your email and responded there! For anyone else that gets this issue just make sure you add all the variables into your project. If some variables are not defined within your Voiceflow project, but you call on them in javascript blocks, things won't work.

I think there is an error in your code. This is the code from the vid: // Extract input variables from args const { delay } = args.inputVars; // Validate that the required input variable is provided if (!delay) { return { next: { path: 'error' }, trace: [{ type: "debug", payload: { message: "Missing required input variable: delay" } }] }; } // Define the URL for fetching user information, incorporating the userId const url = `httpbin.org/delay/${delay}`; // Configure the fetch request const config = { method: 'GET', headers: { 'Content-Type': 'application/json', }, }; try { // Make the API call const response = await fetch(url, config); // Check if the response status is OK if (!response.ok) { throw new Error(`HTTP error! status: ${response.status}`); } // Create the success return object with extracted data return { next: { path: 'success' }, trace: [ { type: "text", payload: { message: `Delayed for ${delay} seconds.` } } ] }; } catch (error) { return { next: { path: 'error' }, trace: [{ type: "debug", payload: { message: "Error: " + error.message } }] }; } -->Paste the code in between this part of your RAW function export default async function main(args) { **PASTE HERE** }

@@BartSlodyczka Awesome thank you!

Hey Aaron, will do! I'll make videos + share templates for "AI autocorrect" and "summarise ticket" in the next couple of days 💪

Thanks man! here you go creator.voiceflow.com/dashboard?import=659cefa59c2d1babbf19f139

I'll make this video soon but this should get you started: ruclips.net/video/wAG4KPOSJkQ/видео.htmlsi=9pJIk03-BfepBz0d

Tutorial here: ruclips.net/video/rznlqKgGg3A/видео.htmlsi=YHjfbvyW-cwlm3hz

Hey legend, if you also want the google sheets part pls send me a DM on discord or email to bart@supportlaunchpad.com. I split them up cos most ppl just wanted the shopify stuff

@@BartSlodyczka I've messaged you on discord!

Here is an MVP I built that should get you started: ruclips.net/video/wAG4KPOSJkQ/видео.htmlsi=9pJIk03-BfepBz0d

@abdelkrimsaida-zs3tc yeah you need to create an api key in shopify, and then plug that api key into the API call

@@BartSlodyczka which plan do I need to get access to the shopify api? does it work with the basic plan?

@@benthinze1606 I think all shopify plans give access to generate an api key