I'm currently working on an integration between Botpress and Zenvia, which is our WhatsApp-based customer support tool (similar to Zendesk). The bot is already running and we’ve successfully connected it to Zenvia for sending and receiving messages. However, we are trying to implement a HITL (Human-In-The-Loop) system, where a human agent can take over a conversation and later hand it back to the bot. The problem is: messages sent directly by human agents through Zenvia are not recognized or processed by Botpress as part of the conversation context. In other words, the bot continues operating but does not "see" or consider the human-sent messages when generating its next responses. What we've noticed/tried so far: The bot works perfectly in fully automated flows. When a human sends a message, it appears in the conversations tab within the dashboard, but Botpress does not include it in the dialogue history. We haven't found any official method to inject those human messages into the Botpress context retroactively. My question: ➡️ Is there any way (official or workaround) to feed human-sent messages from Zenvia into the Botpress context so that the bot can continue conversations seamlessly after HITL interventions? Any API or strategy to manually append to the conversation history? Appreciate any ideas, insights, or examples you can share. Thanks in advance! 🙏 https://cdn.discordapp.com/attachments/1390085199626240172/1390086148369416202/image.png?ex=6866fa1c&is=6865a89c&hm=26a8c7a65a51105034025e7f946d1acee2f5de2203dcbfacc11dca3527f1cbbd&

Hi, I have an autonomous node and it works perfectly in my emulator but doesn't give any output in the webchat? Does anyone know why this could be the case and how to fix it? Thanks, Het

I want to separate out my data by state into multiple KBs, and then have a single Autonomous node be able to use multiple sub workflows, each with a search on their own KB, and compile the data across the different subworkflows and then respond. Currently I am only seeing it choose one subworkflow and then when it returns, the autonomous node answers the whole question based on that one subworkflow. I need it to take a question that applies to multiple states and then make a query for each relevant subworkflow, and then compile all the answers from all the subworkflows. Or, is it better for me to give one KB Search card access to all the state-based KBs and let it do it's thing? Having one KB with all the documents, sometimes the bot gets confused between similar content but in different docs for different states. So my goal is to just have it search KB's individually so there's no chance of confusion there

Hi everyone! Hope you’re all doing well. I haven’t been online for a while and need some help with something. I made a chatbot using Botpress for a web designer. There’s a “My Work” section that shows his websites. I used cards with URLs, so when you click one, it opens the website in a new tab. However, after you come back to the chatbot, that section (the button for that website) disappears. https://cdn.discordapp.com/attachments/1390010569209286780/1390010569372860416/Screenshot_2025-07-02_002511.png?ex=6866b3b8&is=68656238&hm=ff2d85fe5e9e918993ec8a50b8e37d002794cbe0cc8d056bb1f6878f0eb04cad& https://cdn.discordapp.com/attachments/1390010569209286780/1390010569733439488/Screenshot_2025-07-02_002530.png?ex=6866b3b8&is=68656238&hm=2e0a28ec672894a53ed8ec9e36147bd4d23ad386bb1a06d5f9fcab0e4e6ba5a0&

HITL escalation is working fine but all chats are going into the same chat on FreshChat. There's no way to tell who's who when multiple chats are escalated. How do you solve this? https://cdn.discordapp.com/attachments/1389997059796369408/1389997060165603459/Screenshot_2025-07-03_at_1.50.37_AM.png?ex=6866a724&is=686555a4&hm=88cbe04d68cf37668f3fa5c21cc368006aa3993ac1675ea7375c82aad5b40221&

Hie i have been trying to use the execute code editor for the whole day. I type in a promt click enter but then nothing happens after that even if i paste a piece of code there is no execute button showing . I have tried to use different browsers to login , restared my computer cleared the catch of the browser but nothing ia working kindly assit. https://cdn.discordapp.com/attachments/1389973754561433760/1389973755740164136/rn_image_picker_lib_temp_b332ee4d-15d7-4b58-a6de-f21412721c11.jpg?ex=6866916f&is=68653fef&hm=f3a6c2a5c5d96047021d9e83c1569b5ffb15334559ebcd2c0844e369952f6a96&

This code is currently displaying the response on the main card, but since it is currently displaying them individually, the response is split into several parts and displayed. How can I display the incoming response on the main card and display it by appending it to the end of the current main card when the response is updated? https://cdn.discordapp.com/attachments/1389924587671650505/1389924588208656455/image.png?ex=686663a5&is=68651225&hm=ed5e39805912984131324a1ea7ca47790a4de74cde2b6229882a5f8ffe6efb91&

Hi, I'm working on a chatbot for my team and I'm trying to integrate with Sharepoints. I got all credentials from my IT. My problem relates (I think 👀) to the name of the sharepoints database. I followed the Botpress <> Sharepoints integration video, at 4:24 you're mentioning where to take the Document Library Name from. It seems like in my case, it should be "Shared%20Documents" (with full link being: https://DOMAIN.sharepoint.com/sites/StoreKnowledgebase/Shared%20Documents/Forms/AllItems.aspx). I tried "SharedDocuments" and some other combinations too. When this is filled in, I get the registration error (attached). When I leave it empty "[]", my configuration is saved without the error but the Knowledge Base remains empty. I don't know if this is something I'm doing wrong - if so, any tips? Or is this a configuration that my IT should be fixing. What do you think? https://cdn.discordapp.com/attachments/1389900818261610588/1389900818983026788/image.png?ex=68664d82&is=6864fc02&hm=4522009e9ff7f61d61efe156302cbadafc5282e5e83a256f310cbda25fc650a2&

We started using Botpress by manually creating and configuring bots in Studio. But now we want to make it more production like. We need to track changes, store current configuration in version control, etc. I think that can be achieved by either: * using API to store / write workflows. * defining bot as a code https://botpress.com/docs/for-developers/sdk/bot/getting-started But I have no luck getting anything working. How are you managing changes in the bots?

40 users, 400 messages, 750 LLM calls, 3 average bot messages per session - 2 of these are Hi(s) and Thankyou(s) - Using GPT 4o (Nov), got $11 in less than 2 days? Can I somehow check these LLM calls? Also, any chance the agi/improvement plugin is the culprit? (Because I know agi/connor uses a lot of table quota for some weird reason) https://cdn.discordapp.com/attachments/1389852932345692201/1389852932866048091/image.png?ex=686620e9&is=6864cf69&hm=98a0be46319698e6bbfe2ea79a275a1882de8edba86becaca38859eb9d66f079&

Does anyone know how I can change the email address for my BotPress Cloud account?

Hey, I'm facing some trouble that hasn't happen until today I identified within the logs of an integration of mine. The error shown are indexed, what it means? How can i fix that? https://cdn.discordapp.com/attachments/1389734528905384057/1389734529123749991/image.png?ex=6865b2a3&is=68646123&hm=48afda21a8e7eb99d2d9583c4318b2aa2071a0335db157daecd2fa516761a9e0&

When sending a Webchat Custom Trigger event from the browser, the workflow does not execute from the trigger if the workflow is already on a different node. Screenshots have the simple workflow and two test cases with actual and expected results (expected results are in the console comments). Is this a bug or expected behavior? I've spent hours fighting with this problem so any help is greatly appreciated. https://cdn.discordapp.com/attachments/1389717441604485131/1389717442644934816/image.png?ex=6865a2ba&is=6864513a&hm=af715dabcb1cbdc0a8e40a39371243b4931c0c9f3c548289f119de7580de43ba& https://cdn.discordapp.com/attachments/1389717441604485131/1389717443122827334/image.png?ex=6865a2ba&is=6864513a&hm=3474673419c3bb2f0934da3b456e1e193b5dba8350a306c76a03e97ba9d20ea8& https://cdn.discordapp.com/attachments/1389717441604485131/1389717443567554720/image.png?ex=6865a2ba&is=6864513a&hm=730e8d622ef8972f2551d73d70a2e6c66b01a14a8f0a3b986d839bcc1df862ec&

Hi everyone, I am trying to integrate zapier within an automation I have built on botpress but I am struggling to make it work. Is there anyone that can help me with this please?

I'm building a chatbot using botpress and I need to handle response using streaming because of reduce response time. so just DM me. if you resolve this, I will hire you for big project. just DM me

Hi everyone, I’m trying to speed up image-content extraction in Botpress by running the Vision Agent on multiple images at once. Right now I’m using an Autonomous Node with the built-in “Extract from Incoming Images” card, but it processes each extractImageContent call in series and sometimes hits the 15 s timeout when I feed it more than a few images. What I’ve tried so far: - Providing the Autonomous Node with example scripts (IN THE PROMPT, NOT AN ACTION) for it to replicate (didn’t speed it up) - Giving the Autonomous Node very specific instructions on how to loop and return results (no improvement) Also, I’m evaluating whether a custom Action could help. Specifically, I’d like to execute something like this inside an Action (ideally, using a

Promise.all()

to process in parallel instead of series):

// The user sent an image, so extract its content, save it, and transition
const imageUrl = "https://files.bpcontent.cloud/2025/07/01/10/20250701103803-BL3K25XM.png";
const imageDescription = await visionAgent.extractImageContent({ imageUrl });
workflow.description = (workflow.description || "") +
  "\n===IMAGE_BREAK===\n" + imageDescription;
workflow.imageUrls = [imageUrl];
workflow.urlsDone = true;
workflow.transition({
  reason:    "you have saved the description",
  condition: "you have saved the description"
});

Questions: 1) Can Botpress Actions invoke await visionAgent.extractImageContent(...) the same way Autonomous Nodes do? 2) If not, are there hidden config flags or internal endpoints to enable this? Any other tips for offloading image extraction into Actions or speeding it up? Thanks in advance! https://cdn.discordapp.com/attachments/1389557349680676935/1389557351325110353/Screenshot_2025-07-01_at_11.44.11.png?ex=68650da1&is=6863bc21&hm=9d3e34311edb444b28ac3a1894ad8c3c4073d59cf1530354ad7087d520c7bade&

Hi everyone, I'm running into a frustrating issue when trying to import a .bpz bot into the latest Botpress Studio (Web version). ✅ What I did: Created a new flow ("BeautySalonFlow") and defined it cleanly in JSON, added it properly to the bot.json (id, name, startNode, etc.), zipped everything into a .bpz and uploaded it. 😩 The problem: The flow doesn’t show up in the Studio, even though it’s clearly listed in the bot.json. I even tried manually creating the flow in the Studio first with the same name – no effect. 👉 My question: Which files (besides bot.json) are required for a flow to be recognized and displayed after importing a .bpz? Are there hidden dependencies (like metadata or internal Studio IDs) we need to be aware of? Thanks for your help – I’ve tried every route (manual, bp deploy, zip import) but seem to be stuck on some structural detail 🙃

Hi! 👋 Is there any way to access or view the number of tokens used per conversation in Botpress? I’d like to know how much each conversation costs in terms of tokens or estimated usage. I can see tokens per message in the logs, but I’m wondering if there's a built-in way to track or export total tokens per session (either in the Studio UI or via an API). Thanks in advance!

Hi. I'm looking for a way to limit the number of user interactions in the chatbot. I'm using a autonomous node. I've tried some code I found here, but it doesn't work with the autonomous node. The node repetition limit doesn't work either. I'm hoping anyone can come up with a way to limit this. Thanks.

If anyone has some tips on this I would be very appreciative 🥹 The instructions to pass user data from my app to my bot here do not work for me: https://botpress.com/docs/webchat/interact/send-user-data I have read through lots of threads on this issue on here and it seems like a lot of people have a similar problem. I have tried many of the proposed solutions and still can't get it working. I am firstly trying to see if the user data being sent with window.botpress.init appears in the "event" object. I can see placeholders for user data in several places, e.g. - user.userData - "variable_values": { "userData": {} } - conversation.customData I have tried setting multiple of these variables but nothing changes in the raw "event" object. I have also tried matching up the variables in the code to the variables set inside the botpress UI (bottom left corner). Any help would be much appreciate!! Example of the code I am trying to use

<script>
    const firstName = "John";
    const lastName = "Doe";
    const email = "john.doe@example.com";

    window.botpress.init({
        "botId": "",
        "configuration": {
            "website": {},
            "email": {},
            "phone": {},
            "termsOfService": {},
            "privacyPolicy": {},
            "variant": "soft",
            "themeMode": "light",
            "fontFamily": "inter"
        },
        "clientId": "",
        "user": {
            "userData": {
                "firstName": firstName,
                "lastName": lastName,
                "email": email
            },
            "data": {
                "firstName": firstName,
                "lastName": lastName,
                "email": email
            }
          }
    });

report_01JZ2FT1T2J0FFVNC25E2WHSVM I couldn't find any solutions so finally created a report. @prehistoric-airplane-85682 please have someone look into this when possible. Or should I contact the in-studio Support?

how can a trigger a workflow of ConversationEnd from a hook

Hi all! I’m trying to set the border-radius of the Botpress webchat widget to 0 on mobile devices, but I can’t get it to work. Here’s what I’ve tried: * I added the following to the “Styles” section in the Botpress webchat admin:
‪ ‪ @media (max-width: 767px) {‬
‪    .bpContainer {‬
‪      border-radius: 0 !important;‬
‪    }‬
‪  }‬ * I also tried increasing specificity (e.g., .bpReset .bpContainer), and using border-radius: 0px !important;. * However, the chat window still has rounded corners on mobile. * Inspecting the widget, I see the border-radius is being set via a CSS variable or a stylesheet that loads after my custom styles, so my override isn’t taking effect. Screenshots: (I’ve attached screenshots showing my custom styles and the live widget with the border-radius still present.) Question: Is there a way to force the border-radius to 0 on mobile? Is there a recommended way to override the default styles, or a config option I’m missing? If the border-radius is set via inline styles or CSS variables, is there a workaround? Thanks in advance for any help! https://cdn.discordapp.com/attachments/1389484189962731540/1389484190151610498/Screenshot_2025-07-01_at_12.46.10_PM.jpg?ex=6864c97e&is=686377fe&hm=59554fbd854cb678201e2d6411e8dd1e57e3c2aa2458377fa2e23ff041fc508e& https://cdn.discordapp.com/attachments/1389484189962731540/1389484190428303483/Screenshot_2025-07-01_at_12.46.26_PM.jpg?ex=6864c97e&is=686377fe&hm=650a7ead855fed18ff82c9f2c71b86071af2166d7eb7bcdd3a46bca65894efeb&

Looking to set up FreshChat HITL however, i'm consistently getting

Plugin action execution failed

with the main error being

Error executing card "card:Start HITL" in flow:Main:node:HITL - zs: Failed to get user by email Request failed with status code 400

. Any idea on how to solve this? https://cdn.discordapp.com/attachments/1389483486674681916/1389483488532496505/Screenshot_2025-07-01_at_3.47.38_PM.png?ex=6864c8d7&is=68637757&hm=f0066c9b69b664b74ba420e58369ac37d02804e710f19303751f71fea938b546&

Greetings, everyone. Is it possible to link the WhatsApp number provided by Botpress to a Facebook Page, so that the Call-to-Action (CTA) button in my ads can trigger the bot?

Anyone knows how to use this from instagram integration? https://cdn.discordapp.com/attachments/1389413498777243648/1389413499029164122/image.png?ex=686487a8&is=68633628&hm=eefbb29eee1760c62720889386ca7b000e43cd339b4ce6e733e05c098fb61e19&

Hello, I’m currently using the Autonomous mode in Botpress for my Instagram integration, and I’ve encountered a problem where the bot responds immediately to multiple incoming messages instead of waiting and batching them. Here’s the behavior I expect: If a user sends multiple messages (e.g., "Hi", then "Hi" again within 30 seconds), the bot should wait 30 seconds before replying once, using all received messages in the context. Here’s what is actually happening: The bot responds immediately to each message, sometimes within 6–7 seconds. I tried creating a custom hook with logic to delay or skip processing if the last message was recent, but it didn’t work reliably. I also couldn’t insert custom code inside the Autonomous1 node directly. Even after trying the "Buffering Messages and Delaying Response" card or setting delays, the bot still replies twice or more instead of once. My goal is: To delay the response for 30 seconds from the last incoming message, and then respond only once using the collected inputs. Can you please guide me on how to properly handle this in Autonomous mode? Or is there a recommended pattern for message buffering and throttling that works reliably in this scenario?

I'm trying to embed a Botpress chatbot inline on my website using the method in the official docs (window.botpress.init() with a container). I've correctly set the botId, clientId, and selector, and loaded the required script. The domain is also whitelisted under Allowed Origins. Despite this, the chatbot iframe doesn't render or appears invisible. The bot simply doesn’t show inline as expected.

Is it not possible to upload files on the preview link? What is the easiest way to create a multi-media bot with botpress?

I have a use case where I want the ai chatbot to print out invoices. Can I have the AN ask for the info and the user sends it back in an unstructured text? And the AI would know how to use it like ChatGPT does for example? Or do I always need to use the capture card for such case?