I have been trying to do this in javascript and set a variable but I can't get it to work (with chatGPT's help).

There should be an easy way to pass the url of the uploaded image and get the width dimension right?

I also need to do it to get the height dimension.

How would this be done?

I tried it like this:

https://mydomain.com/?MyArray=["007BFF","007BFF","00FF05","00FF05","00FFFB","C800FF","C800FF","FF0000","FF0000","FF9500","FF9500","FFEA00","FFEA00","FFEA00","FFFFFF"]

And it just loads that as the text string, when I try to access the array with

{{={{MyArray}}[0]}}=}} etc, I just get

[
"
0

Just the characters in the string

I tried it even encoding the query to

%5B%22007BFF%22%2C%22007BFF%22%2C%2200FF05%22%2C%2200FF05%22%2C%2200FFFB%22%2C%22C800FF%22%2C%22C800FF%22%2C%22FF0000%22%2C%22FF0000%22%2C%22FF9500%22%2C%22FF9500%22%2C%22FFEA00%22%2C%22FFEA00%22%2C%22FFEA00%22%2C%22FFFFFF%22%5D


But I get the same result.

Is it possible to load a variable with an array/list via the query string, or would I need to instead send a reference to a table/db and then load the array/list in the bot? (that would just make it all more cumbersome).

Thanks!

So my main bot pulls down redemption codes from a google sheet, but it has to each use. Is there any way to save the codes inside the bot, like as some sort of universal array and then just increment the pointer when one gets used and on the last value in the array, then it could go to the sheet and grab the next 100 or something codes? This way it only has to query the sheet once every 100 uses instead of every use. The problem I ran into is when a lot of instances are querying at once, it crashes the bot.

Just for further info, the bots are querying the same sheet then putting the used codes back in to avoid duplicates. When the sheet makes new codes, it checks them against the used codes so it doesn't create duplicates. This takes a bit of time for each query and I set a cell in the sheet to busy with the bot for each bot to take and put the code one by one so that the codes don't get double pulled.

So I guess just wondering if there are universal variables or data storage in the bot itself.

If not, I am open to ideas for a more robust solution where all bots can access the same list of available redemption codes and expire them as needed.... Ideas?

I am using randomly generated numbers in a google sheet to create customer/user IDs to track things. My script creates random 10 digit numbers- and it works great BUT when the number in the sheet has leading zeros, and I pull that into typebot, typebot gets it with the leading zero- then when I put it BACK in the sheet, the leading zeros get cut off by default. So it seems typebot will GET as a string but POST to the sheet as a number maybe... (or the sheet interprets it as a number...)

Then when I try to find the value in the sheet to modify row values, it is searching with the identifier with the leading zeros but in the sheet the number doesn't have leading zeros so it never finds the row. I am going to limit my range to only leave non zeros as the first number (start at 1000000000) for the range, but I still think this is kind of a bug. Not sure it needs to be addressed, but maybe an option to get the data as a string or get it is a number. What do you think Baptiste? Not a big deal if this is not a priority. Just found it and wanted to report it.

Videos attached for better explination. Curious about how & why typebot works certain ways & if some improvements can be made.

I have learned that if there is number of things to be done in the background, typebot does it all, THEN prints to the screen all at once and AFTER it finishes the processing on the back end along with displaying the request for user input.

This is annoying because I sometimes I want to give the user a headsup such as "Ok, now processing your entry, this may take a mintue so please be patient" - but I'll put that in BEFORE a long flow of tasks to do. It process and hang only to send that message to the user seconds or a minute later and right before it asks the user for input.

Could there be a way to SIMULATE a user input- even if it doesn't do anything?

Flow could process in segments a message to user such as:

"Image sucessfully updated"
"Image processsing"
"Creating Image choices"

These breaks could also give a change for evaluations and giving the customer notification if something failed.

I made two videos, and the second one it DOES send info to the user.

I am attaching 2 videos that show both things happening. One, the data is all processed and logged but it takes a minute or two to print text to the screen for each loop (testing text) but I am curious as to why, if the processing is done in the log, does the output text take so long to output to the user instead of just quickly as the result is already processed...What is the differnce? Sheet blocks? The video will make this make sense.

So along with this, I am wondering if it would be possible to put in markers/simulated user input, if we could somehow use typebot as a nocode automator.... I am going to make feature requests...

Video one - log is done but screen output takes FOREVER
https://d.pr/v/gx6C4C

Video two - Outputs text as it goes with no customer input. WHY? Google sheets checks?
https://d.pr/v/ws0gms

I needed to apend array values in a variable into another variable holding the array of arrays. It saves correctly but when I access on of the arrays and try to use it in a function, it does not work- it seems to evaluate as a string.

I am going to try to just save them to nocodb and access them back from there and see if that works, BUT I think this may be a bug.

To recreate it, just generate random arrays of values like 3, then save those arrays into another array. Then access one of those arrays by index and try to process them in a function.

@Baptiste -If needed I coudl set up an example of this but I'm just super busy. But if you need me to throw that together, let me know!

Does anyone know how to get an image URL from an AI?

I don't really see a good option in the openAI block to get the image URL back.

I've tried conveting image blobs/strings before but it doesn't work great for me and I don't see a way to return that either but maybe I am misssing something.

Now, I have not learned to train an assistant yet so maybe that is the path I need to go down. But also, maybe there is something I am missing with openAI or maybe one of the other integrations gives me good image generation?

Thanks for any help!

I love nocoDB, it is so fast.

But I just noticed a bug and it seems significant-

If there are no matches for the "Select Records" logic in the block, instead of not writing any data to the database, it writes the update to EVERY cell in the column. This seems like a HUGE bug. I can put the update inside of typebot logic but that will just make it all more complex since it is NOT working in the NocoDB block.

It shows 9:30 AM August 15th, is that in my timezone or @Baptiste 's timezone?

Just want to confirm as I want to try to attend these when possible.

Ok, so I found a weird bug.

I wanted to display a number that was one higher than the current number (as an option for the minimum amount a user can enter to increase the number)

Intead of precalculating that number and storing in another variable before the text, I used the inline format of {{={{MyVar}}+1=}}.

Later in the text I display {{MyVar}}

Well, it displays the value for {{MyVar}} about 6 characters too early. So for example:

"This is the increased value {{={{MyVar}}+1=}} and then here is the starting value {{MyVar}}."

Let's say the value was 12 this would display as:

"This is the increased value 13 and then here is the 12value {{MyVar}}."

I made a bot to show how it was behaving in the code I was creating and then decided to also just put in this simple example since I think I know what is happening- (sample bot URL in first comment).

I think typebot will adjust the index by 4 to account for the opening and closing {{CurlyBrackets}} on any variable. BUT when there are EXTRA symbols when an inline function is used, typebot is not set up to account for those and so the {{=Inline curly brackets and equals signs=}} throw off the display on the final string of any variables that come after the inline code by 6 because they were removed but the index wasn't adjusted for those with regards to inserting the variables...

Hope I explained that correctly. I was trying to figure out what was going on as I slept and in the shower this morning and I think this is it.

Check it @Baptiste and let me know when it is fixed should be a quick one.

Is there a way to access these saved variables from within the bot?

I want to just check them for duplicates to see if a redemption code has been issued before (Also storing them in an external database)

I would only need to query like the last minute or less (but open to querying more depending on posibilities). Once a code is generated and no duplicates are found, I do store it in the database as a used code so I just need to query the bot for codes created but haven't made it to the database yet.

Ok, looking for documentation or specific help. I want to use imagga.com api to analyze an image and then give me back the most common color or colors.

Here is the documetation:

https://docs.imagga.com/?shell#colors

There are different code examples on the right. But I just need to know where to put things in the typebot block setup.

How do I set up the athentication and then how do I put in the parameters? To make it all work.

I will be using the upload URL for the uploaded images and just neeed the API to analyze and return colors as HEX or even RGB values.

Thanks!

In general, any good help on understanding and using APIs would be good too. I do want to learn and get away from no-code tools such as Zapier as much as possible if I can just use the APIs myself.

Ok, so this is one that probably isn't THAT important but seems to be a bug nonetheless.

If I have button module saving to a variable, and one of the choices is:

TRUE

And one is

True

Then no matter which I select, it saves in the variable as the Capialized version. But if there is only one choice spelled the same it will take the case of that button text. If there is one button with more capitalized letters, it will take the option with the most capitalized letters and place it in the variable value.

So there would be limited reasons why this would be done, and I discovered it testing/thinking about something else, but it does appear to be a minor bug as All caps vs no caps/some caps should be able to be used as separate variable value posibilities.

In my bot I send all of the variables to some links and sheets but I account for all of them- I don't understand the utility of the "Merge Answers" option. Can someone explain it to me- I looked in the documentation too.

Like in what case does toggling it on or off affect anything?

Thanks!

This is a question concerning an ongoing discussion.

Baptiste, you will be happy that I am simplifying my (CRAZY) large bots. So I am breaking them up into smaller organized ones.

So the issue that I am having-

Bot 1 (Variable is created and assigned value) > Bot 2 (Variable is created) = Works fine

Bot 1 (Variable is created and assigned value) > Bot 2 (Variable is not used, not created, doesn’t exist) > Bot 3 (Variable is created) = Doesn’t work the Variable does not have the assigned value from Bot 1.

The issue is I have a crazy amount of customized variables and to sit and create ALL of the variables in ALL of the simple bots takes a LOT of time just to simplify out to the different bots (versus doing them all in one). Is this how it is supposed to work, or should these variables be saved to the session and passed between bots and when I need them I can create the variable in that bot and use the value?

My understanding before is they should persist and be available to me without me having to create a TON of variables in every bot even if that bot doesn’t use that variable. Please let me know thanks!