Documentation

Dashboard

The Dashboard is your central command center. It provides a high-level overview of your chatbot's knowledge base, usage statistics, and provides the tools you need to deploy the chatbot on your website.

From here, you can monitor key metrics and quickly access the necessary scripts for integration.

Dashboard Analytics

The top of the dashboard features several key metrics to help you understand the scale of your chatbot's knowledge and its usage.

Statistics Overview

These boxes provide a quick snapshot of your chatbot's data sources and overall token consumption.

• Number of Files: Shows the total count of documents and images you have uploaded to train the chatbot.
• No. of Website Links: Displays the total count of website URLs that have been successfully scraped and are part of the chatbot's knowledge base.
• Total Words: A sum of all the words from all your data sources (documents, websites, etc.).
• Total Used Tokens: Shows the cumulative number of tokens consumed by the chatbot in conversations. This includes both the user's input (Input Tokens) and the chatbot's response (Output Tokens).

Used Tokens Chart

This chart provides a visual representation of your token usage over time. It helps you track consumption patterns and understand when your chatbot is most active.

You can change the view to see data aggregated by different time periods:

• By Session: Shows token usage for each individual conversation.
• Daily: Aggregates all token usage for each day.
• Monthly: Sums up token usage for each month.
• Yearly: Provides a total for each year.

Deploying Your Chatbot

This section provides the necessary code and links to integrate your chatbot into your website or use it directly.

Installation Script

To embed the chatbot on your website, you need to add a small snippet of code to your site's HTML. This code loads the chatbot's interface and connects it to your account.

1. Copy the entire code block provided in the text area using the "Copy" button.
2. Paste this code just before the closing tag in your website's HTML file(s).
3. Once you save and publish your website, the chatbot icon will appear, ready to interact with your visitors.

Other Deployment Options

• Open Chatbot: Click this button to open your chatbot in a new, dedicated browser tab. This is useful for testing or sharing a direct link to your chatbot.
• Download Wordpress Plugin: If your website is built on WordPress, you can use this button to download a dedicated plugin for easy, no-code installation.

Live Preview

The dashboard includes a live preview of your chatbot. You can use this interface to test its responses and behavior in real-time, based on its current training data, without needing to install it on your live website first.

Website

The Website data source section is the central hub for managing all content scraped from web pages. You can add content in several ways and then manage it through its lifecycle: from unscraped, to trained, to archived.

Add Sitemap

A sitemap is the most efficient way to add all your website's pages at once. It's a file that lists every page on your site, ensuring the chatbot can discover and learn from all available content.

How to Add a Sitemap:

1. From the main Scraper page, click the "Add Sitemap" button.
2. A modal window will appear.
3. Enter the full URL of your sitemap (e.g., https://yourcompany.com/sitemap.xml).
4. Click the "Add Sitemaps" button.

The system will then add all the URLs found in the sitemap to the "Unscraped Data" queue for processing.

Add Homepage

If you don't have a sitemap, you can add your website's homepage. The system will crawl this page and automatically discover and add other pages that are linked from it.

How to Add a Homepage:

1. From the main Scraper page, click the "Add Homepage" button.
2. In the modal window, enter your full homepage URL (e.g., https://yourcompany.com).
3. Click the "Add Homepage" button.

The URLs found by the crawler will be added to the "Unscraped Data" queue.

Add Website Links

This option allows you to add individual web page URLs one by one. This is useful for adding specific pages that may not be in your sitemap or for quickly adding new content.

How to Add a Single Link:

1. From the main Scraper page, click the "Add Website Links" button.
2. In the modal window, enter a single, full URL of a web page.
3. Click the "Add Website Link" button. The URL will be added to the "Unscraped Data" queue.

List Sitemap

You can view and manage the sitemaps you have previously added.

How to View Added Sitemaps:

1. From the main Scraper page, click the "List Sitemap" button.
2. A side panel will open, showing a list of all sitemap URLs you have submitted.
3. You can permanently delete a sitemap from this list by clicking the trash icon next to it.

Unscraped Data

This is the first tab in the Scraper section. It acts as a queue, listing all the website links that have been added but have not yet been processed (scraped) for content. From here, you initiate the training process.

How to Use This Table:

1. Use the checkboxes on the left to select the URLs you want to process. You can select up to 99 links at a time.
2. Once you have made your selection, a set of buttons will appear.
3. Click the "Scrape Data and Train" button to begin processing the selected links.
4. A status modal will appear, showing the real-time progress of each URL being scraped.

Successfully scraped links will be moved to the "Trained Data" tab, while any links that fail will be moved to the "Error Data" tab.

You can also delete links from this queue individually with the icon or in bulk with the "Bulk Delete" button.

Error Data

This tab contains a list of all website links that failed during the scraping process. This could be due to the page being unavailable, blocked, or in an unreadable format.

Managing Errors:

• Review the Error: The table includes a column with the specific error message to help you diagnose the issue.
• Retry Scraping: You can select one or more failed items and click the "Scrape Data and Train" button to try processing them again.
• Delete: You can remove links from this list permanently using the individual or bulk delete actions.

Trained Data

This is the core knowledge base of your chatbot. It contains all the website data that has been successfully scraped, processed, and is actively being used to answer questions.

Understanding the Table:

• URL/Page Name: Identifies the source of the data.
• Auto Sync & Sync Schedule: Controls if and how often the system automatically checks the live webpage for content changes. Click on these fields to configure them.
• Status: Shows if a page has a "Pending" update that needs your review.

Managing Trained Data:

You can perform actions on individual items using the "more" icon () at the end of each row, or on multiple items by selecting them with the checkboxes.

Key Actions:

• Edit Content: Opens a rich-text editor allowing you to manually modify the raw text content that was scraped from the page. This is useful for removing unwanted sections (like ads or menus).
• Edit Summary: Opens an editor to modify the AI-generated summary of the content. This summary is often used by the AI for quicker responses.
• Refresh & Retrain: Re-scrapes the live webpage to fetch the latest content and retrains the chatbot on it.
• Archive: Removes the data from the active knowledge base and moves it to the "Archive Data" tab. The chatbot will no longer use this information.
• Delete: Permanently deletes the data from the system.

Pending Updates:

If Auto-Sync is enabled, the system may find changes on a webpage. When this happens, the status will change to "Pending". Click this status to open a comparison view where you can review the differences and either Approve the new content or Reject the changes.

Archive Data

This tab holds data that you have removed from the active training set but have not permanently deleted. The chatbot does not use information from the archive.

Managing Archived Data:

From here, you have two main options for each item:

• Retrain & Unarchive (): This action moves the data back to the "Trained Data" tab, making it active again in the chatbot's knowledge base.
• Delete Permanently (): This action will permanently remove the data from the system. This cannot be undone.

These actions can be performed on individual items or in bulk by selecting multiple rows.

Documents

In addition to website content, you can train your chatbot by directly uploading files or creating text documents. This is ideal for internal knowledge bases, product manuals, policy documents, or any information not publicly available on your website.

Upload or Create Document

The "Manage Documents" page is your starting point. From here, you can either upload existing files from your computer or create a new document from scratch using a text editor.

How to Upload Files:

1. Click the file input area or drag and drop your files into it. You can select multiple files at once.
2. Supported file types include: .docx, .pdf, .xlsx, .pptx, .json, .xml, .txt, .md.
3. (Optional) If the uploaded document originates from a specific webpage, you can enter its URL in the "Document Source" field.
4. Click the "Upload File(s)" button.

Once uploaded, your documents will be processed and added to the "List Documents" table below.

Create New Document (Text Editor)

If your content doesn't exist in a file, you can create it directly within the application using a rich-text editor.

How to Create a New Document:

1. On the "Manage Documents" page, click the "Create New Document (Text Editor)" button.
2. A modal window will appear with a text editor.
3. Enter a Filename for your document. The system will automatically add the .docx extension.
4. (Optional) Add a "Source Website Link" if the content is related to a specific URL.
5. Type or paste your content into the editor. You can use the toolbar to format your text with headings, lists, bold, italics, links, and tables.
6. Click the "Save Document" button to create the file and add it to your knowledge base.

List Documents

This table displays all the documents you have uploaded or created. It is the central place for managing your document-based knowledge.

Understanding the Table:

• File Name: The name of the uploaded or created document.
• Page Content/Summary: A brief preview of the document's content.
• Is Trained?: Indicates if the chatbot has been trained on this document ("Yes" or "No").
• Upload Date: The date the document was added to the system.
• Action: A set of icons for managing the document.

You can use the search bar at the top right to quickly find specific documents.

Edit Document

You can modify the content of any document after it has been uploaded. This is useful for correcting typos, updating information, or removing irrelevant text.

How to Edit a Document:

1. In the "List Documents" table, find the document you want to edit.
2. Click the pencil icon in the "Action" column for that row.
3. A modal window will open with the document's content loaded into the text editor.
4. Make your desired changes to the text or the "Source Website Link".
5. Click the "Save" button to update the document.

Edit Summary

The system automatically creates a summary of each document. You can edit this summary to refine the key information the chatbot uses for quick responses.

How to Edit a Summary:

1. In the "List Documents" table, find the relevant document.
2. Click the summary icon in the "Action" column.
3. A modal window will appear, loading the summary text into an editor.
4. Modify the summary as needed.
5. Click the "Save Summary" button to apply your changes.

Delete Document

If a document is no longer needed, you can permanently remove it from the system.

How to Delete Documents:

• Single Deletion: Click the trash icon in the "Action" column of the document you wish to remove. Confirm the action when prompted.
• Bulk Deletion: Use the checkboxes on the left side of the table to select multiple documents. An action bar will appear; click the "Bulk Delete" button and confirm the action.

Warning: Deleting a document is permanent and cannot be undone.

Retrain Chatbot

After you upload, create, edit, or delete documents, you must retrain the chatbot for the changes to take effect.

The "Retrain Chatbot" button is located at the top of the "List Documents" table. Clicking this button tells the AI to re-read all of its available knowledge (from both documents and websites) and update its understanding.

How to Retrain:

1. Navigate to the "List Documents" tab.
2. Click the "Retrain Chatbot" button.
3. The button will display a "Training..." status.
4. Once complete, the button text will change to "Chatbot Trained", and a success message will appear. Your chatbot is now updated with the latest information.

Image

Train your chatbot to understand the content within images. By uploading images, the system uses Optical Character Recognition (OCR) to extract text, which can then be used as part of the chatbot's knowledge base. This is useful for training on infographics, scanned documents, diagrams, and more.

Add Image

The "Manage Images" page allows you to upload image files directly from your computer. The system will process these images to extract any readable text.

How to Upload an Image:

1. Click the file input area to open a file selection window, or drag and drop your image files onto it.
2. Select one or more images from your computer. Supported file types are .jpg, .jpeg, and .png.
3. Click the "Upload Files" button. A spinner will indicate that the files are being uploaded and processed.

Upon successful upload, a confirmation message will appear, and the uploaded images will be added to the "List Images" table below.

List Images

This table displays all the images you have uploaded. From here, you can review the extracted text and manage each image entry.

Understanding the Table Columns:

• File Name: The original filename of the uploaded image.
• Page Content: A preview of the text that was extracted from the image using OCR.
• Is Trained?: Shows "Yes" if the chatbot has been trained on this image's content, and "No" otherwise.
• Upload Date: The date the image was uploaded.
• Action: Provides icons to edit or delete the image data.

Edit Image Data

After an image is processed, you can edit the extracted OCR text and its summary. This is crucial for correcting any OCR errors or refining the information the chatbot learns.

How to Edit Image Data:

1. In the "List Images" table, find the image you wish to edit.
2. Click the pencil icon in the "Action" column for that row.
3. A modal window will open with two main sections:
   • Image OCR Content: A rich-text editor containing the text extracted from the image. You can correct any mistakes, format text, and remove irrelevant information here.
   • Summary: A text box where you can write or edit a concise summary of the image's content. This helps the AI quickly understand the context of the image.
4. After making your changes, click the "Save" button at the bottom of the modal.

Delete Image

If an image is no longer relevant or was uploaded by mistake, you can permanently delete it from the system.

How to Delete an Image:

1. Find the image you want to remove in the "List Images" table.
2. Click the trash icon in the "Action" column.
3. A confirmation prompt will appear. Confirm the action to permanently delete the image and its associated data.

Warning: Deleting an image is permanent and cannot be undone.

Retrain Chatbot

After uploading new images or editing the data of existing ones, you must retrain the chatbot for the changes to be applied to its knowledge base.

The "Retrain Chatbot" button is located at the top of the "List Images" table. Clicking this button initiates a process where the AI re-learns from all available data sources, including your newly updated image content.

How to Retrain:

1. After making your changes, navigate to the "List Images" tab.
2. Click the "Retrain Chatbot" button.
3. The button's text will change to "Training..." to indicate the process has started.
4. Once finished, a success message will appear. Your chatbot is now updated with the latest image data.

Audio

Train your chatbot by recording your voice directly in the browser. The system will transcribe your speech into text, which is then added to the chatbot's knowledge base. This is an excellent way to add conversational Q&A, spoken instructions, or any expert knowledge you can articulate.

Record Audio

The "Audio Transcription" interface allows you to easily record, transcribe, and add audio data to your chatbot.

How to Record Audio:

1. Select Language: Before recording, choose the language you will be speaking from the dropdown menu. This is essential for accurate transcription.
2. Start Recording: Click the "Start" button. Your browser may ask for permission to use your microphone. Please allow it.
3. Speak Clearly: A timer will begin, showing the duration of your recording. Speak clearly into your microphone.
4. Stop Recording: When you are finished, click the "Stop" button.

After you stop the recording, the system will automatically upload the audio and begin the transcription process. A "Transcribing..." indicator will appear. Once complete, a success message will be displayed, and your new audio entry will appear in the "List Audio" table below.

Note: There is a maximum recording time available based on your subscription plan. This limit is displayed below the recording buttons.

List Audio

This table displays all your processed audio recordings. Each row represents a single recording that has been transcribed and is now part of the chatbot's knowledge base.

Understanding the Table:

• File Name: A unique name assigned to the audio recording file.
• Page Content: A preview of the text transcribed from your audio.
• Is Trained?: Indicates if the chatbot has been trained on this transcription.
• Upload Date: The date and time the recording was made.
• Action: Provides icons to edit or delete the transcription.

You can use the search bar at the top right to quickly find specific recordings by searching their transcribed content.

Edit Audio

Automated transcription is powerful, but not always perfect. This feature allows you to edit the transcribed text to correct any errors, ensuring the chatbot learns from accurate information.

How to Edit a Transcription:

1. In the "List Audio" table, find the recording you wish to correct.
2. Click the pencil icon in the "Action" column for that row.
3. A modal window will open with the transcribed text loaded into a rich-text editor.
4. Make your desired changes to the text. You can use the toolbar for basic formatting.
5. Click the "Save" button to update the transcription.

Delete Audio

You can permanently remove any audio recording and its transcription from the system.

How to Delete an Audio Record:

1. Find the audio recording you want to remove in the "List Audio" table.
2. Click the trash icon in the "Action" column.
3. A confirmation prompt will appear. Confirm the action to permanently delete the recording.

Warning: This action is permanent and cannot be undone.

Retrain Chatbot

After you record new audio, edit a transcription, or delete a recording, you must retrain the chatbot. This critical step updates the AI's memory with your latest changes.

The "Retrain Chatbot" button is located at the top of the "List Audio" table. Clicking this button tells the AI to re-read all of its knowledge sources (including websites, documents, and audio transcriptions) and rebuild its understanding.

How to Retrain:

1. Navigate to the "List Audio" table after making your changes.
2. Click the "Retrain Chatbot" button.
3. The button will show a "Training..." status while it works.
4. When the process is complete, the button will update to "Chatbot Trained," and a success message will appear. Your chatbot is now ready with the updated information.

Knowledge

The Knowledge section provides a centralized view of all the information your chatbot has learned. It combines the data from all your sources—websites, documents, images, and audio—into a single, searchable knowledge base. This is the master list of everything your chatbot knows.

List Knowledge

The "Knowledge Data" table displays every piece of content that has been successfully processed and is available for the chatbot to use. This unified view helps you understand the full scope of your chatbot's knowledge at a glance.

Understanding the Table:

• Data: Shows the source of the information, which will be either a website URL or a filename (for documents, images, and audio).
• Content: A preview of the text content extracted from the source.
• Data Type: A color-coded label that clearly identifies the source type (e.g., website_link, document, image, voice). This helps you quickly distinguish between different kinds of data.
• Date: The date the data was created or uploaded.
• Actions: Icons that allow you to edit or delete the knowledge item.

You can use the search bar at the top of the table to filter the entire knowledge base and find specific information quickly.

Edit Knowledge

You have the ability to edit the text content of any item in your knowledge base, no matter its original source. This is perfect for making final corrections, refining information, or removing irrelevant text to improve the chatbot's accuracy.

How to Edit Knowledge:

1. In the "Knowledge Data" table, find the item you want to modify.
2. Click the pencil icon in the "Actions" column for that row.
3. A modal window will appear, loading the item's text content into a rich-text editor.
4. Make your desired changes. You can use the toolbar for formatting like bold, lists, and headings.
5. Click the "Save" button to update the information in the chatbot's knowledge base.

Delete Knowledge

If a piece of information is outdated, irrelevant, or incorrect, you can permanently remove it from the chatbot's knowledge base from this central view.

How to Delete a Knowledge Item:

1. Find the item you want to remove in the "Knowledge Data" table.
2. Click the trash icon in the "Actions" column.
3. A browser confirmation prompt will ask you to confirm the deletion.

Warning: Deleting an item from the knowledge base is a permanent action and cannot be undone. The chatbot will immediately lose access to this information.

Connectors

Connectors allow you to link your chatbot to external services like Google Drive. By establishing a connection, you can easily sync and train your chatbot on documents, spreadsheets, and other files stored in the cloud, keeping its knowledge base up-to-date automatically.

Google Drive

The Google Drive connector lets you securely link your account to select files and folders for chatbot training. This is a powerful way to manage your knowledge base using the familiar interface of Google Drive.

Connect to Google Drive

To begin, you must first authorize the application to access your Google Drive files.

How to Connect:

1. Navigate to the Google Drive Connector page.
2. Click the "Connect to Google Drive" button.
3. You will be redirected to a Google authentication page. If you are not already logged in, you will be prompted to sign in to your Google account.
4. Google will ask you to grant permission for the application to access your files. Review the permissions and click "Allow" or "Continue".
5. After granting permission, you will be redirected back to the application. The page will now show that you are connected and will display a table for managing your synced files.

List & Manage Files

Once connected, this page will display a table of all the files you have synced from your Google Drive. This is your central hub for managing your cloud-based knowledge source.

Managing Your Synced Files

The table provides several options for managing each file:

• Scrape & Train: After syncing new files, select them using the checkboxes and click the "Scrape & Train Selected" button. This action processes the file's content and trains the chatbot on it.
• Edit Content (): Opens a modal with a text editor, allowing you to modify the text extracted from the file. This is useful for correcting errors or removing irrelevant content.
• Refresh (): Re-fetches the latest version of the file from Google Drive and retrains the chatbot on the new content.
• View in Drive (): Opens the file directly in Google Drive in a new browser tab.
• Delete (): Removes the file from your synced list and its content from the chatbot's knowledge. This does not delete the file from your Google Drive.

Auto Sync and Pending Updates

You can configure files to sync automatically. Click on the status under the "Auto Sync" or "Schedule" columns to open the sync settings modal.

• Sync Settings: In the modal, you can enable automatic syncing and set a custom interval (in days) for how often the system should check for updates.
• Review Pending Updates: If auto-sync is on and the system detects a change in a file you've previously edited manually, the status will change to "Review". Clicking this opens a side-by-side comparison, allowing you to edit and accept the new version or reject it to keep your manually edited version.

Select Files to Sync

After connecting your account, the first step is to choose which files or folders the chatbot should learn from. This is done using the Google Drive Picker.

How to Sync Files and Folders:

1. On the Google Drive Connector page, click the "Select Files/Folders to Sync" button.
2. The Google Drive Picker window will open, showing the files and folders in your drive.
3. Navigate through your drive and select one or more files or entire folders.
4. Click the "Select" or "Open" button in the picker window.
5. The selected files will be added to the sync list in the application. You will see them appear in the main table on the page.

Next Step: Scrape and Train

After syncing files, they are listed but not yet part of the chatbot's knowledge. You must select the new files from the table and click the "Scrape & Train Selected" button to complete the process.

Disconnect

If you no longer want the application to have access to your Google Drive, you can revoke its permissions at any time.

How to Disconnect:

1. On the Google Drive Connector page, click the "Disconnect" button.
2. Confirm the action if prompted.

This will securely remove the application's authorization token. All automatic syncing will stop, and you will need to reconnect to add or refresh any more files.

OneDrive

The OneDrive connector lets you securely link your Microsoft account to select files and folders for chatbot training. This is a powerful way to manage your knowledge base using the familiar interface of OneDrive.

Connect to OneDrive

To begin, you must first authorize the application to access your OneDrive files. This process is secure and handled directly by Microsoft.

How to Connect:

1. Navigate to the OneDrive Connector page.
2. Click the "Connect to OneDrive" button.
3. You will be redirected to a Microsoft authentication page. If you are not already logged in, you will be prompted to sign in to your Microsoft account.
4. Microsoft will ask you to grant permission for the application to access your files. Review the permissions and click "Accept" or "Continue".
5. After granting permission, you will be redirected back to the application. The page will now show that you are connected and will display a table for managing your synced files.

Note: Only users with an "Admin" role can initiate the connection to OneDrive. Other users will see a message to contact their administrator.

List & Manage Files

Once connected, this page will display a table of all the files you have synced from your OneDrive. This is your central hub for managing your cloud-based knowledge source.

Managing Your Synced Files

The table provides several options for managing each file:

• Scrape & Train: After syncing new files, select them using the checkboxes and click the "Scrape & Train Selected" button. This action processes the file's content and trains the chatbot on it.
• Edit Content (): Opens a modal with a text editor, allowing you to modify the text extracted from the file. This is useful for correcting errors or removing irrelevant content. This does not change the original file in your OneDrive.
• Refresh (): Re-fetches the latest version of the file from OneDrive and retrains the chatbot on the new content.
• View in OneDrive (): Opens the file directly in OneDrive in a new browser tab.
• Delete (): Removes the file from your synced list and its content from the chatbot's knowledge. This does not delete the file from your OneDrive.

Auto Sync and Pending Updates

You can configure files to sync automatically. Click on the status under the "Auto Sync" or "Schedule" columns to open the sync settings modal.

• Sync Settings: In the modal, you can enable automatic syncing and set a custom interval (in days) for how often the system should check for updates.
• Review Pending Updates: If auto-sync is on and the system detects a change in a file you've previously edited manually, the status will change to "Review". Clicking this opens a side-by-side comparison, allowing you to edit and accept the new version or reject it to keep your manually edited version.

Select Files to Sync

After connecting your account, the first step is to choose which files the chatbot should learn from. This is done using a dedicated OneDrive file browser.

How to Sync Files:

1. On the OneDrive Connector page, click the "Select Files from OneDrive" button.
2. A new popup window will open, showing the files and folders in your OneDrive.
3. Navigate through your folders, select one or more files using the checkboxes, and then click the "Add files" button.
4. The selected files will be added to the sync list in the application. You will see them appear in the main table on the page.

Next Step: Scrape and Train

After syncing files, they are listed but not yet part of the chatbot's knowledge. You must select the new files from the table and click the "Scrape & Train Selected" button to complete the process.

Disconnect

If you no longer want the application to have access to your OneDrive, you can revoke its permissions at any time. This option is available to users with an "Admin" role.

How to Disconnect:

1. On the OneDrive Connector page, click the "Disconnect" button.
2. Confirm the action if prompted.

This will securely remove the application's authorization. All automatic syncing will stop, and you will need to reconnect to add or refresh any more files.

Channels

Channels are the different platforms where you can deploy your chatbot to interact with your users. This section guides you through connecting your chatbot to popular messaging services like WhatsApp.

Whatsapp Bot

Integrate your chatbot directly with your WhatsApp Business account. This allows your chatbot to automatically handle customer conversations, answer questions, and provide support 24/7 on one of the world's most popular messaging platforms.

Whatsapp Bot Integration

To connect your chatbot to WhatsApp, you need to provide credentials from your Meta (Facebook) for Developers account. This page provides you with the necessary `Callback URL` and `Verify Token` and requires you to input your unique WhatsApp API keys.

Configuration Steps

Follow these steps carefully to link your chatbot to WhatsApp.

Step 1: Get Your Credentials from Meta for Developers

You will need a Meta Developer account and a configured WhatsApp Business App. You can set this up at developers.facebook.com. Once your app is set up, find the following information:

1. WhatsApp Access Token: In your app's dashboard, go to the "WhatsApp" > "API Setup" section. You will find a temporary access token here. For production use, you should generate a permanent one through the Business Manager.
2. WhatsApp Phone Number ID: This ID is located directly below the access token in the "API Setup" section. It's the unique identifier for the phone number you are using for the bot.
3. WhatsApp Business Account ID: This can be found in your Meta Business Settings under "Business Info".
4. WhatsApp Number: This is your business phone number that is registered with the WhatsApp Business API, including the country code (e.g., 15551234567).
5. WhatsApp Verify Token: This is a token you create yourself. It acts like a password between your chatbot and WhatsApp. Enter a secure, random string of text (e.g., "my_secret_token_123") in this field. You will need to use this exact same token in the Meta dashboard in the next step.

Step 2: Configure Webhooks in Your Meta App

You now need to tell Meta where to send WhatsApp messages. In your Meta App's "API Setup" dashboard, find the "Webhook configuration" section and click "Edit".

1. Callback URL: On our application's integration page, you will see a "Callback URL". Copy this entire URL.
2. Paste this URL into the "Callback URL" field in your Meta dashboard.
3. Verify Token: Copy the "Verify Token" you created in the previous step and paste it into the "Verify Token" field in your Meta dashboard.
4. Click "Verify and Save". Meta will then send a test request to our application to ensure the connection is valid.
5. After verifying, click "Manage" under webhooks and subscribe to the "messages" field.

Step 3: Save Configuration

Once you have configured the webhook in Meta, return to our application's integration page.

1. Fill in all the fields with the tokens and IDs you gathered in Step 1.
2. Click the "Connect with WhatsApp" button.

If the connection is successful, you will see a confirmation message. Your chatbot is now connected to WhatsApp! If you see an error, double-check that all tokens and IDs have been copied correctly and that your access token has not expired.

Agents

AI Agents are specialized tools that give your chatbot new abilities beyond just answering questions from your knowledge base. They can connect to external services like Google Calendar or a MySQL database to perform actions, such as booking meetings or fetching live data. This section explains how to configure and manage these powerful agents.

The main "Manage Your AI Agents" page provides an overview of all available agents, their current status, and links to configure them.

Understanding the Agent Card

• Configured / Not Configured: This status tells you if the necessary credentials and settings have been saved for the agent. You must configure an agent before you can activate it.
• Active / Inactive Toggle: This switch allows you to enable or disable an agent. An agent must be configured before it can be activated. When active, the chatbot can use its special abilities. When inactive, it will behave as a standard Q&A bot.
• Configure / Manage Configuration Button: This button takes you to the detailed setup page for the specific agent.

Google Calendar Agent

The Google Calendar Agent allows your chatbot to interact directly with your Google Calendar. Users can ask the chatbot to schedule new events, check for upcoming meetings, or ask questions about their schedule, all through natural conversation.

Configuration Steps

The setup process is guided by a stepper that walks you through connecting your account and defining the agent's behavior.

Step 1: Connection Details

First, you must authorize the agent to access your Google Calendar.

1. Click the "Connect to Google Calendar" button.
2. You will be redirected to a Google sign-in page. Log in and grant the requested permissions.
3. After authorizing, you will be redirected back to the agent configuration page, which will now show a "Connected" status.

You can use the "Disconnect" button at any time to revoke access.

Step 2: Agent Backstory

This is where you define the agent's personality and role. It's a set of instructions that tells the AI how it should think of itself. A detailed default backstory is provided, which you can customize to fit your needs.

Step 3: Task Description

This section contains detailed instructions on how the agent should perform its tasks. You can define how it should interpret user requests, how to handle dates and times, and which tools to use for reading versus creating events. A comprehensive default description is provided, which is recommended for most users.

Saving and Activating the Agent

1. After completing all three steps, click the "Save Configuration" button at the bottom.
2. Once the configuration is saved, you can enable the agent using the "Active / Inactive" toggle switch at the top of the page.

MYSQL Agent

The MySQL Agent enables your chatbot to connect to your MySQL database and perform read-only (SELECT) queries. This allows users to ask questions and receive live data directly from your database, such as checking product stock, looking up order statuses, or getting user information.

Security Note: For security, it is highly recommended to create a dedicated, read-only database user for the chatbot and to only grant it access to the specific tables and columns it needs.

Configuration Steps

Follow the stepper to connect your database and instruct the agent on how to use it.

Step 1: Database Connection Details

Provide the credentials for your MySQL database. All fields are required.

• Database Host: The IP address or domain name of your database server (e.g., localhost or db.yourcompany.com).
• Database Name: The name of the specific database the agent should connect to.
• Database User: The username for the read-only database account.
• Database Password: The password for the database user.

Step 2: Agent Backstory

Define the agent's role and context. Critically, you should specify the tables and columns the agent is allowed to query. A default backstory is provided as a template.

Step 3: Task Description

Provide explicit instructions for the agent on how to translate natural language into SQL queries. You should explain the schema (tables and columns) and give examples of how to handle different types of user questions. A detailed default description is provided that you can customize with your specific table and column names.

Saving and Activating the Agent

1. After providing the database credentials and reviewing the agent's instructions, click the "Save Configuration" button.
2. The system will test the connection. If successful, the agent will be marked as "Configured".
3. You can then enable the agent using the "Active / Inactive" toggle switch at the top of the page.

Gmail Email Agent

The Gmail Email Agent connects directly to your Google Account, empowering your chatbot to manage your emails. It can read, search, count, and even reply to emails based on your instructions, turning your chatbot into a powerful email assistant.

Agent Status

Before configuring the agent, you can control its overall state.

• Active/Inactive Switch: This toggle allows you to turn the agent on or off. The agent must be connected to your Gmail account before it can be activated. When inactive, it will not process any email-related requests.
• Edit Router Agent: This button opens settings to configure how the main "Router Agent" directs tasks to this specific Gmail agent.

How to Configure the Gmail Agent

Configuration is a simple 3-step process that guides you through connecting your account and defining the agent's behavior.

Step 1: Connection Details

This is the first and most crucial step. Here you will authorize the application to access your Gmail account.

• To Connect: Click the "Connect to Gmail" button. You will be redirected to a Google sign-in page to securely grant the necessary permissions. Once done, you'll be returned to this page, and the status will show as connected.
• To Disconnect: If you are already connected, a "Disconnect" button will be visible. Clicking this will revoke access and deactivate the agent.

You cannot proceed to the next steps until the agent is successfully connected.

Step 2: Agent Backstory

In this step, you define the agent's personality and background. This "backstory" gives the AI context on its role, helping it behave more consistently.

A default backstory is provided, describing the agent as an expert email assistant. You can modify this text to better suit your needs or leave it as is for optimal performance.

Step 3: Task Description

Here, you provide a detailed set of instructions that the agent must follow when handling email tasks. This description tells the agent which tools to use for specific keywords (e.g., use the "List Emails" tool when a user asks to "find" an email).

A comprehensive default description is already included, covering common scenarios like searching, counting, and replying to emails. You can customize these instructions to create more specific workflows.

Saving the Configuration

After you have completed the connection and are satisfied with the backstory and task description, click the "Save Configuration" button at the end of the process. This will save your settings and ensure the agent behaves according to your definitions.

Outlook Email Agent

The Outlook Email Agent integrates your chatbot with your Microsoft Outlook account. Once connected, it can perform a variety of email-related tasks, such as reading, searching, and replying to messages, effectively transforming your chatbot into a smart assistant for your Outlook inbox.

Agent Status

This section at the top of the page allows you to manage the agent's operational state.

• Active/Inactive Switch: Use this toggle to activate or deactivate the agent. The agent cannot be activated until it is successfully connected to your Outlook account. When inactive, it will not handle any email-related functions.
• Edit Router Agent: Click this button to adjust how the main "Router Agent" prioritizes and sends tasks to the Outlook agent.

How to Configure the Outlook Agent

Setting up the agent is a straightforward process divided into three simple steps, guiding you from authentication to defining the agent's behavior.

Step 1: Connection Details

This initial step involves authorizing the agent to access your Outlook account securely.

• To Connect: Click the "Connect to Outlook" button. You will be redirected to a Microsoft login page to grant the required permissions. After successful authorization, you will return to this page with a "Connected" status.
• To Disconnect: If an account is already linked, a "Disconnect" button will be displayed. Clicking it will revoke the agent's access to your Outlook account and automatically deactivate it.

You must complete this connection step before you can proceed with the configuration.

Step 2: Agent Backstory

Here, you can define the agent's background and personality. This "backstory" provides context to the AI, helping it understand its role and interact more effectively.

A default backstory is provided, casting the agent as an expert in email management. You can customize this text or use the default for reliable performance.

Step 3: Task Description

This step allows you to provide specific instructions that the agent will follow when performing email tasks. The description outlines how the agent should use its tools in response to user requests (e.g., use the "Count Emails" tool when a user asks "how many").

A comprehensive default template is included to handle common tasks like searching, reading, and replying to emails. You can modify these rules to create custom workflows.

Saving the Configuration

Once you have connected your account and are satisfied with the agent's backstory and task description, click the "Save Configuration" button. This will apply all your settings, and the agent will be ready to manage your Outlook emails.

Campaigns

The Campaigns module allows you to proactively reach out to your customers on various channels. You can send marketing messages, announcements, or notifications to a list of contacts, all managed from a single place.

From the main Campaigns page, you can select which type of campaign you would like to configure and run.

Whatsapp Bulk Messaging

The WhatsApp Bulk Messaging campaign allows you to send personalized messages to a list of contacts from a CSV file. The process is guided by a simple step-by-step wizard to ensure your campaign is set up correctly before sending.

Step 1: Upload a Recipient File

The first step is to provide a list of contacts to message. This is done by uploading a CSV (Comma-Separated Values) file. This file should contain a header row with column names, such as `Name` and `whatsapp_number`.

• Drag and drop your CSV file into the designated area or click "Browse files" to select it from your computer.
• Once the file is selected, click the "Upload CSV" button to process the file and move to the next step.
• Skip File Upload: If you have previously uploaded a list of recipients and want to reuse it, you can click this button to proceed directly to the recipient selection step without uploading a new file.

Step 2: Select Recipients

In this step, you will see a table of all the contacts from your uploaded CSV file. You can choose which contacts should receive your message.

• Use the checkboxes on the left to select individual recipients.
• Click the "Select All" checkbox in the table header to select all contacts in the list.
• Once you have made your selection, click the "Next" button to proceed.

Step 3: Create Your Message Template

Here, you will write the message that will be sent to your selected contacts. You can also save your message as a template for future use.

• Template Name: Give your message template a memorable name (e.g., "Monthly Newsletter" or "Special Offer").
• Message Content: Write your message in the main text area.
• Personalization: You can personalize your message by using variables that match the column headers in your CSV file. Wrap the column name in curly braces `{}`. For example, if you have a column named `Name`, you can write `Hello {Name},` in your template. When the message is sent, `{Name}` will be automatically replaced with the contact's actual name from the file.
• Save Template: Click the "Save Template" button to save the current name and message for easy reuse in future campaigns.

After crafting your message, click "Next".

Step 4: Review and Send

This is the final step before your campaign is sent. Review the details to ensure everything is correct.

• The screen will show you the total number of recipients you have selected and a preview of your message template.
• When you are ready, click the "Send Bulk Messages" button to launch the campaign.

Monitoring Send Status

After you click send, the screen will update to show the status of each message being sent. You can see which messages were delivered successfully and which ones failed, allowing you to track your campaign's performance in real-time.

Marketing

The Marketing module provides tools to promote your products or services directly through your chatbot. This section covers how to create, manage, and display advertisements to your users during their conversations.

Campaigns

Campaigns are specific marketing initiatives you can run. Currently, this includes managing advertisements that your chatbot can display to users.

Gmail Auto-Responder

The Gmail Auto-Responder transforms your chatbot into a smart email assistant. By connecting to your Gmail account, it can automatically read incoming emails, use its AI to generate intelligent replies, and even be trained on past conversations to improve its responses over time. This allows you to automate customer support, answer common inquiries, and manage your inbox more efficiently.

Connect to Gmail

To get started, you must first securely authorize the application to access your Gmail account. This process is handled by Google's secure authentication system.

How to Connect:

1. Navigate to the Gmail Auto-Responder page from the "Campaigns" menu.
2. Click the "Connect to Gmail" button.
3. You will be redirected to a Google sign-in page. Choose the account you wish to connect and sign in.
4. Google will ask for permission to allow the application to read and send emails on your behalf. Review the permissions and grant access.
5. After successful authorization, you will be returned to the application, which will now display the main email management interface.

Note: Only users with an "Admin" role can connect or disconnect the Gmail account for the team.

How to Disconnect:

To revoke access, simply click the "Disconnect Gmail" button located at the top of the page and confirm the action. This will stop all email processing immediately.

Managing Incoming Emails

Once connected, the main screen displays a table of all incoming emails that the system is processing. This is your command center for reviewing email status, managing replies, and training the AI.

Understanding Email Status

Each email in the table has a status badge to help you quickly understand its current state:

• REPLY GENERATED: The AI has read the email and successfully drafted a reply, which is now waiting for your review and approval to be sent.
• REPLIED AI: An AI-generated reply has been sent for this email.
• REPLIED MANUAL: A manually edited reply has been sent for this email.
• REPLIED (TRAINED): A reply has been sent, and this conversation has been used to train the AI to improve future responses.
• PROCESSING: The system is currently analyzing the email to generate a reply.
• ERROR: An error occurred while processing the email.

Filtering and Fetching Emails

• Filters: At the top of the table, you can use the dropdown filters to narrow down the list by "Training Status" (e.g., show only emails the AI has been trained on) or by "Email Status" (e.g., show only emails with a reply ready to be sent).
• Check for New Emails: The system checks for new emails periodically. To check immediately, click the "Check for New Emails" button.

Replying to Emails

You have full control over when and how replies are sent. You can review and edit each reply individually or send approved replies in bulk.

Reviewing and Sending a Single Reply

To review an individual email and its AI-drafted reply, click the view icon in the "Actions" column. This opens the Reply Modal.

• Conversation History (Left): This panel shows the entire email thread for context, so you can understand the full conversation.
• Your Reply (AI Draft) (Right): This is a rich-text editor pre-filled with the AI-generated reply. You can make any changes you want—correcting details, changing the tone, or adding more information.
• Send Reply: Once you are satisfied with the reply, click the "Send Reply" button. The message will be sent from your connected Gmail account.

Sending Replies in Bulk

For efficiency, you can send multiple approved AI-generated replies at once.

1. In the main email table, use the checkboxes to select the emails that have a "REPLY GENERATED" status.
2. Click the "Bulk Reply" button that appears at the top of the table.
3. The system will send the pre-generated AI replies for all the emails you selected.

Training the AI

You can improve the AI's performance by training it on successful past conversations. When you "train" on an email, you are telling the AI that the sent reply was a good example, helping it generate better, more accurate replies in the future.

Individual Training

You can train (or untrain) the AI on any email that has already been replied to.

• To Train: For an email that has been replied to but is not yet trained, an action icon to "Train AI" will be available. Click this to use the conversation as a positive example for the AI.
• To Untrain: If an email has already been used for training, you will see an "Untrain" icon . Clicking this will remove its influence from the AI's knowledge.

Bulk Training

You can also train the AI on multiple conversations at once.

1. In the main email table, select the emails that have a "REPLIED" status but are not yet trained.
2. Click the "Bulk Train" button at the top of the table.
3. The AI will learn from all the selected conversations, improving its future performance.

Configuring AI Settings

You can customize the core behavior of the AI by editing its master prompt and enabling automatic replies. To access these settings, click the "Settings" button at the top of the email list.

AI Prompt Template

This is the most important setting for controlling the AI's personality and behavior. The prompt is a set of instructions that guides the AI on how to write its replies. You can define its tone (e.g., friendly, professional), its role (e.g., customer support agent), and how it should structure its response.

You can use placeholders like {sender_name} and {email_body} in the prompt to provide the AI with context from the actual email it is replying to.

Enable Automated AI Replies

This toggle switch controls whether the AI sends replies automatically.

• When disabled (default): The AI will generate a reply and wait for your manual approval. The email's status will be "REPLY GENERATED".
• When enabled: The AI will automatically send the reply it generates without waiting for manual review. The email's status will go directly to "REPLIED AI".

Use the automated replies feature with caution, and ensure your prompt is well-defined to prevent unintended responses.

After making changes, click the "Save Settings" button to apply them.

Outlook Auto-Responder

The Outlook Auto-Responder integrates your chatbot with your Microsoft Outlook account, turning it into an intelligent email assistant. It automatically reads incoming emails, drafts smart replies using AI, and can be trained on past interactions to continuously improve its accuracy. This feature helps you automate customer support and manage your inbox with greater efficiency.

Connect to Outlook

To begin, you must securely authorize the application to access your Outlook account. This process is managed by Microsoft's secure authentication system.

How to Connect:

1. Navigate to the Outlook Auto-Responder page from the "Campaigns" menu.
2. Click the "Connect to Outlook" button.
3. You will be redirected to a Microsoft sign-in page. Choose the account you want to connect and sign in.
4. Microsoft will ask for permission for the application to read and send emails on your behalf. Review the permissions and grant access.
5. Upon successful authorization, you will be returned to the application, which will now display the email management interface.

Note: Only users with an "Admin" role can connect or disconnect the Outlook account for the team.

How to Disconnect:

To revoke access, click the "Disconnect Outlook" button at the top of the page. This will immediately stop all email processing and securely remove the connection.

Managing Incoming Emails

Once connected, the main screen shows a table of all incoming emails being processed by the system. This dashboard is your central hub for reviewing email status, managing replies, and training the AI.

Understanding Email Status

Each email is labeled with a status badge to show its current state:

• REPLY GENERATED: The AI has drafted a reply that is ready for your review and approval.
• REPLIED AI: An AI-generated reply has been sent.
• REPLIED MANUAL: A manually edited reply has been sent.
• REPLIED (TRAINED): A reply has been sent, and this conversation was used to train the AI.
• PROCESSING: The system is actively analyzing the email.
• ERROR: An error occurred during processing.

Filtering and Fetching Emails

• Filters: Use the dropdown menus at the top of the table to filter the list by "Training Status" or "Email Status".
• Check for New Emails: While the system checks for emails periodically, you can click the "Check for New Emails" button to trigger an immediate check.

Replying to Emails

You have complete control over how replies are sent. You can review each one individually or send multiple approved replies in bulk.

Reviewing and Sending a Single Reply

To review an email and its AI-drafted reply, click the view icon in the "Actions" column. This opens the Reply Modal.

• Conversation History (Left): Shows the entire email thread for full context.
• Your Reply (AI Draft) (Right): A rich-text editor pre-filled with the AI's reply. You can edit the content as needed before sending.
• Send Reply: When you're ready, click the "Send Reply" button. The message will be sent from your connected Outlook account.

Sending Replies in Bulk

To save time, you can send multiple AI-generated replies at once.

1. In the email table, use the checkboxes to select emails with the "REPLY GENERATED" status.
2. Click the "Bulk Reply" button located above the table.
3. The system will send the pre-generated AI replies for all selected emails.

Training the AI

Improve the AI's future performance by training it on successful past conversations. When you train on an email, you are teaching the AI that the sent reply was a good example, helping it generate higher-quality replies over time.

Individual Training

You can train (or untrain) the AI on any email that has already been replied to.

• To Train: For a replied email that is not yet trained, click the "Train AI" icon to use it as a learning example for the AI.
• To Untrain: For an email already used for training, click the "Untrain" icon to remove its influence.

Bulk Training

You can also train the AI on multiple conversations simultaneously.

1. In the email table, select the emails that have a "REPLIED" status but are not yet trained.
2. Click the "Bulk Train" button above the table.
3. The AI will learn from all selected conversations.

Configuring AI Settings

Customize the AI's core behavior by editing its master prompt and managing automated replies. Access these options by clicking the "Settings" button at the top of the email list.

AI Prompt Template

This is the primary control for the AI's personality and response style. The prompt contains instructions that guide the AI on how to compose replies, including its tone, role, and structure. You can use placeholders like {sender_name} and {email_body} to give the AI context from the email it is replying to.

Enable Automated AI Replies

This setting determines whether the AI sends replies automatically.

• Disabled (Default): The AI drafts a reply and waits for your manual approval (status becomes "REPLY GENERATED").
• Enabled: The AI automatically sends the reply it generates without waiting for review (status becomes "REPLIED AI").

Use automation with care. Ensure your prompt is well-defined to avoid sending unintended replies.

Click the "Save Settings" button to apply any changes.

SMS Management

The SMS Management feature allows you to send text messages directly to your customers' mobile phones. You can send individual messages, run bulk campaigns to a list of contacts, create reusable message templates, and track the delivery status of every message you send.

Note: To use this feature, an administrator must first configure the SMS service provider in the system settings. If the service is not configured, all sending functions will be disabled.

Send SMS

The "Send SMS" tab is your main dashboard for creating and launching SMS campaigns. From here, you can choose your recipients, compose your message, and send it.

How to Send an SMS Message

Click the "Send SMS" button to open the sending dialog. The process involves two main steps:

Step 1: Add Recipients

In the "Phone Numbers" text box, you can add your recipients in two ways:

• Manual Entry: Type or paste phone numbers directly into the box. You must enter one full phone number per line, including the country code (e.g., +1234567890).
• Import from CSV: Use the "Import Contacts" feature (explained below) to automatically populate this list from a file.

Step 2: Compose Your Message

You have two options for writing your message:

• Use a Template: Select a pre-written message from the "Use Template" dropdown menu. This will automatically fill the "Message" box with the template's content.
• Write a Custom Message: Type your message directly into the "Message" text box. Even if you select a template, you can still edit the content here before sending.

Step 3: Send the Message

Once you have added recipients and composed your message, click the "Send SMS" button. The message will be sent to all the phone numbers in your list.

Import Contacts from a CSV File

For large campaigns, you can easily import a list of phone numbers from a CSV file. This saves you from having to enter each number manually.

How to Import Contacts:

1. From the "Send SMS" tab, click the "Import Contacts" button. This will open the import dialog.
2. Prepare your CSV file. It must contain a header row with a column named exactly `phone_number`.
3. Click inside the dialog to select your CSV file, or drag and drop it into the window.
4. Click the "Process CSV" button. The system will read the file and show you a preview of the phone numbers it found.
5. If the preview is correct, click the "Use These Numbers" button. You will be taken back to the "Send SMS" dialog, and the "Phone Numbers" field will be automatically filled with the imported contacts.

Manage Templates

The "Templates" tab allows you to create, edit, and delete reusable SMS messages. This is perfect for frequently sent messages like appointment reminders, promotions, or welcome texts.

How to Create a New Template

1. Click the "Create Template" button.
2. In the dialog that appears, give your template a descriptive Template Name (e.g., "Monthly Promotion").
3. In the Message Content box, write the message you want to save.
4. Click "Save". Your new template will now appear in the list and be available in the "Send SMS" dialog.

Editing and Deleting Templates

To manage your existing templates:

• Edit: Click the pencil icon next to a template to open the editor and make changes.
• Delete: Click the trash icon to permanently remove a template.

Message History

The "Messages" tab provides a complete log of every SMS message sent from the system. Here you can track the delivery status and review the content of past campaigns.

Understanding the History Table

• Phone Number: The recipient's mobile number.
• Message Preview: The first part of the message that was sent.
• Status: The delivery status of the message, indicated by a colored chip:
  • PENDING: The message is in the queue to be sent.
  • SENT: The message has been successfully sent from our system to the carrier.
  • DELIVERED: Confirmation has been received that the message was delivered to the recipient's phone.
  • FAILED: The message could not be delivered (e.g., the number was invalid).
• Sent At: The date and time the message was sent.

You can use the pagination controls at the bottom of the table to navigate through older message records.

Advertisements

This page is your central hub for managing all advertisements that can be displayed by your chatbot. You can create new ads, edit existing ones, and see a summary of all your active and inactive campaigns.

Managing Advertisements

The main table provides a comprehensive overview of your ads:

• Title: The name of your advertisement.
• URL: The target web page users will be sent to if they click the ad.
• Featured: Shows "Yes" if this is a featured ad, which may give it priority placement.
• Days: The number of days the ad is set to run from its upload date.
• Expires / Uploaded: The expiration and creation dates for the ad.
• Actions: Buttons to edit or delete an advertisement.

Add or Edit an Advertisement

Whether you're creating a new ad or editing an existing one, you'll use a simple form to provide the necessary details. To create a new ad, click the "Add Advertisement" button. To edit, click the pencil icon in the actions column of an existing ad.

Form Fields:

• Title: A short, descriptive title for your ad campaign (e.g., "Summer Sale" or "New Product Launch").
• Display Duration: How long the advertisement should be active. You can choose from 7, 15, or 30 days. The expiration date is calculated automatically.
• Advertisement Details: The main text of your ad. This is the message that will be shown to the user (e.g., "Get 20% off all summer products!").
• Target URL: The full website link (starting with http:// or https://) where users will be redirected when they interact with the ad.
• Set as Featured Advertisement: A toggle switch to mark the ad as "featured." You can have multiple featured ads, which may be shown more frequently or in more prominent positions.

After filling out the form, click "Add Advertisement" or "Save Changes" to apply your settings.

Deleting an Advertisement

To permanently remove an advertisement, click the trash icon in the actions column of the corresponding ad. A confirmation window will appear to prevent accidental deletion. This action cannot be undone.

Marketplace Settings

This page allows you to configure how your chatbot is presented in the public KlonIT Marketplace. You can create a public listing for your chatbot, set a price for your services, and allow other users to sign up and use your AI-powered expertise.

Publishing Details

This section controls the public-facing information for your chatbot's listing in the marketplace. Fill out these details to attract users and clearly explain what your chatbot does.

Configuration Fields

• Publish Chatbot to Marketplace: This is the master switch for your listing. When toggled on, your chatbot will be visible to everyone in the marketplace. When off, it remains private.
• Chatbot Title: A short, catchy name for your chatbot that will be displayed as the main title of your listing. This is a required field.
• Profession Category: Select the category that best describes your chatbot's area of expertise (e.g., "Software Development," "Legal Services"). This helps users find your chatbot when browsing the marketplace.
• Short Description: A brief summary of your chatbot's purpose and capabilities, limited to 300 characters. This appears in list views and search results. Use the rich-text editor to format your text.
• Long Description: A detailed explanation of your chatbot. Use this space to describe its features, what problems it solves, who it's for, and how to use it effectively.
• Thumbnail Image: Upload an image to represent your chatbot. This will be used as the main visual for your listing. For best results, use a 400x400 pixel image in PNG, JPG, or WEBP format.

Saving Your Settings

After you have filled in all the required details, click the "Save Publishing Settings" button to update your marketplace listing.

Signup Details

This section allows you to monetize your chatbot by setting prices and enabling a signup option on your marketplace page. When enabled, other users can subscribe to and pay for the services your chatbot provides.

Allowing User Signups

The first step is to enable the signup feature.

• Allow users to sign up for this service: Toggle this switch on to reveal the pricing and description fields. When this is active, a "Sign Up" or "Subscribe" button will appear on your public marketplace page.

Pricing and Description

Once you enable signups, you can set your prices and describe what users will get.

• Price Per Hour: Set an hourly rate for your services. This is ideal for consultations, live support, or any service billed by time.
• Price Per Instance: Set a fixed price for a single, one-time task or project.
• Price Per Month: Set a recurring monthly fee. This is suitable for ongoing access, subscription packages, or retainer-based services.
• Service Description: Use the rich-text editor to explain what is included in your service packages. You can describe the scope of work, mention if live agent support is available, and detail your areas of specialization.

Saving Your Details

After configuring your pricing and description, click the "Save Signup Details" button to apply the changes to your marketplace page.

Tracking

The Tracking module is essential for monitoring your chatbot's performance, understanding user interactions, and diagnosing potential issues. Here, you can review conversation histories, check system logs for errors, and analyze direct feedback from your users to continuously improve the chatbot experience.

Sessions

The Sessions page provides a complete, chronologically ordered history of every conversation your chatbot has had with users. This allows you to review interactions, understand user behavior, and identify areas where the chatbot's responses can be improved.

How to Review a Conversation

1. Conversation List: On the left side of the screen, you will find a list of all chat sessions. Each session is identified by a unique ID and the time it occurred (e.g., "5 minutes ago").
2. Select a Session: Click on any session from the list to load the full conversation transcript on the right.
3. View Transcript: The right-hand panel displays the entire dialogue, showing the user's questions and the chatbot's corresponding answers in a familiar chat format.
4. Navigate Pages: Use the pagination controls at the bottom of the session list to browse through older conversations.

Logging

The Logging section is a technical tool for monitoring the health of your chatbot system. It records important system events, with a focus on exceptions and errors. Reviewing these logs can help identify and diagnose technical issues that may arise during the chatbot's operation.

Understanding the Log Table

• Reference ID: A unique identifier for each log entry.
• Log Type: The category of the log (e.g., "Database Error", "API Failure").
• Status: The current status of the logged event.
• Created/Updated On: Timestamps for when the event was logged and last updated.

Viewing Log Details

To see the full technical details of a log entry, simply click on its row in the table. An expanded view will appear below the row, showing the complete error message or log details, which can be useful for troubleshooting.

Feedbacks

This section collects all the direct feedback that users have submitted while interacting with your chatbot. Analyzing user feedback is one of the most effective ways to understand their experience and identify opportunities for improving the chatbot's knowledge and performance.

Reviewing User Feedback

The table displays a summary of each piece of feedback received:

• Questions: The user's question that prompted the feedback.
• Answers: The chatbot's answer that was rated by the user.
• User Issue: The category of the issue reported by the user (e.g., "Incorrect Information", "Unhelpful").
• User Feedback: Any specific comments or notes the user provided.
• Create Date: When the feedback was submitted.

You can use the search bar to filter feedback based on keywords, and click on any row to expand it and view the full details of the feedback entry.

Additional Settings

This section provides advanced tools for fine-tuning your chatbot's conversational abilities. You can improve how it understands user questions, establish definitive "correct" answers for specific topics, and customize its language to better match your brand's voice.

Question Quality

The Question Quality tool allows you to refine the chatbot's understanding of user intent. By providing better-phrased alternatives to real user questions, you can teach the chatbot to recognize different ways a user might ask for the same information. This improves its ability to provide the right answer even when the question isn't phrased perfectly.

How to Improve a Question

1. The main table lists past user questions and the answers the bot provided.
2. Identify a question that could be phrased more clearly or a conversation where the bot struggled.
3. Click the pencil icon in the "Action" column for that row.
4. An editing view will appear, showing the original user question and the bot's answer.
5. In the "Improve Your Question" field, enter a clearer, idealized version of the user's original question. For example, if the user asked "how much stuff is there," you might improve it to "What is the current stock quantity?".
6. Click "Save" to submit your improvement.

By providing these improved examples, you help the chatbot build connections between ambiguous user queries and the high-quality questions that lead to the best answers.

Ground Truth

The Ground Truth feature is a powerful tool for ensuring factual accuracy. It allows you to establish a definitive, "correct" answer for a specific question. When a user asks this question, the chatbot will use your ground truth answer, overriding any other information it might have found in its knowledge base. This is essential for company policies, pricing details, or any information that must be exact.

How to Set a Ground Truth Answer

1. The table shows past conversations. Find an interaction where the bot's answer was incorrect, incomplete, or could be improved.
2. Click the pencil icon in the "Action" column for that row.
3. An editing view will appear, displaying the original question and the bot's answer.
4. In the "Improve Your Answer" field, write the perfect, factual answer that the bot should always provide for this question.
5. Click "Save".

From now on, whenever a user asks a similar question, the chatbot will prioritize delivering your saved ground truth answer.

Pronouns

This section allows you to manage lists of pronouns to customize your chatbot's language and persona. By selecting which pronouns are active, you can fine-tune how the chatbot refers to itself or understands certain terms, aligning it more closely with your brand's voice.

How to Manage Pronoun Lists

The interface consists of two boxes: a list of available items on the left and a list of active items on the right.

1. Select Items: Click on one or more items in either list to select them. You can hold down `Ctrl` (or `Cmd` on Mac) to select multiple individual items.
2. Move Selected Items:
   • Click the single right arrow to move selected items from the left box to the right box (activating them).
   • Click the single left arrow to move selected items from the right box to the left box (deactivating them).
3. Move All Items:
   • Click the double right arrow to move all items from the left to the right.
   • Click the double left arrow to move all items from the right to the left.

The items in the right-hand box are considered "active" and will be used by the chatbot. Make sure to save any changes if a save button is present on the page.