# Bot9 MCP MCP > Bot9 manages your entire conversational AI fleet. This MCP lets you programmatically create, configure, and train custom chatbots for customer support automation. You can manage knowledge sources, track conversation transcripts, and simulate user interactions directly from any agent client. ## Overview - **Category:** customer-support - **Price:** Free - **Tags:** conversational-ai, chatbot-training, knowledge-base, ai-automation, customer-service, agent-orchestration ## Description Need to keep up with multiple bots or constantly feed them new company data? Bot9 connects your AI agents to a centralized system for running and refining conversational workflows. Instead of logging into the bot platform's dashboard, you just use your agent to talk to it—all through natural language. This MCP gives you control over everything: from listing every configured chatbot instance to adding new URLs so they can learn about pricing or policies. You can pull full conversation history for analysis or programmatically send a test message to verify bot responses. Connecting Bot9 via Vinkius means your agent client handles the whole orchestration, giving you one place to manage all of it. ## Tools ### add_data_source Feeds a new URL into the bot's knowledge base so it can learn from that data. ### create_bot Builds and provisions an entirely new, dedicated AI chatbot instance. ### get_bot Pulls the specific configuration details for one bot by its name or ID. ### get_conversation_history Retrieves the complete transcript, including all messages, from a past conversation thread. ### list_bots Lists every chatbot currently configured in your Bot9 account. ### list_conversations Shows the list of current, active conversation sessions for a given bot. ### list_data_sources Displays all knowledge base URLs currently assigned to a specific bot. ### send_message Sends an arbitrary message to the bot and immediately gets back its generated response text. ## Prompt Examples **Prompt:** ``` List all bots in my Bot9 account. ``` **Response:** ``` I've retrieved your bots. You currently have 'Support Assistant' (ID: bot_123) and 'Sales Lead Gen' (ID: bot_456). ``` **Prompt:** ``` Add the URL https://example.com/pricing to bot_123's knowledge base. ``` **Response:** ``` The URL has been successfully added to bot_123. The bot is now training on the new pricing data. ``` **Prompt:** ``` Get the chat history for conversation conv_789 on bot_123. ``` **Response:** ``` Retrieving the history... The user asked 'What are your operating hours?' and the bot replied 'We are open 9AM to 5PM EST'. ``` ## Capabilities ### Manage Chatbot Instances List and retrieve the configuration details for every custom AI chatbot you run. ### Train with Knowledge Sources Add new URLs to the bot's knowledge base, allowing it to learn from updated documentation or websites. ### Monitor Conversations List active chats and pull full transcripts of historical customer interactions for review. ### Test Bot Responses Send messages directly to a bot programmatically, allowing you to test its replies before real users encounter them. ## Use Cases ### Post-Launch Audit A QA lead needs to check if the new 'billing' bot is correctly referencing updated tax rate documentation. They use `list_data_sources` first, then run a targeted `send_message` call to verify the response accuracy. ### Onboarding New Knowledge The technical writing team finishes drafting the new product manual. Instead of manually logging into the bot dashboard, they use their agent to execute `add_data_source`, pointing it straight at the draft's URL. ### Root Cause Analysis A support manager notices a customer complained about a billing error. They ask their agent to retrieve the chat history using `get_conversation_history` for that specific conversation ID, instantly finding where the bot gave outdated information. ### Scaling Chatbot Deployment A product team needs three new specialized bots (Sales, Tech Support, Billing). They use `create_bot` three times in a row to provision all necessary instances without logging into a separate platform. ## Benefits - Stop manually checking bot status. Use `list_bots` to see every chatbot you run in one go, giving you immediate visibility into your whole support stack. - When company policies change, don't wait for a manual update. Run `add_data_source` to feed new URLs directly, keeping the bot informed instantly. - Need to audit how a conversation went? Use `get_conversation_history` to pull full transcripts and analyze exactly what was said on both sides of the chat. - Testing your bots used to mean opening multiple dashboards. Now, you can use `send_message` to simulate user interactions instantly from your agent client. - The ability to list all active chats via `list_conversations` means you never lose sight of a critical support ticket in progress. ## How It Works The bottom line is: you treat bot management like any other function call from your agent client. 1. Subscribe to the MCP and provide your Bot9 API key. 2. Tell your agent what needs doing—like 'list all my bots' or 'add this URL for training'. 3. Your agent executes the action via the MCP, and you get a direct, structured response detailing the result. ## Frequently Asked Questions **How do I start using the `list_bots` tool with Bot9 MCP?** You ask your agent client simply to 'list all my bots.' The MCP returns a list of every bot ID and name you've configured in your account. **Can I use `add_data_source` multiple times for the same bot?** Yes, you can. You run `add_data_source` with a new URL every time your knowledge base content changes, making sure the bot always learns from fresh data. **What is the difference between `list_conversations` and `get_conversation_history`?** `list_conversations` only gives you a list of currently active chats. To read what was actually said, you must use `get_conversation_history` with a specific conversation ID. **Is there a tool to check if the bot is working?** Yes, use the `send_message` tool. You send a test message and immediately get back the bot's generated response, confirming its current status. **How do I handle API key authentication when using the `list_data_sources` tool?** You must provide a valid Bot9 API Key during setup. The system validates this key upon connection; if keys expire or are incorrect, all calls will return an authentication failure error. **Are there rate limits for using the `add_data_source` tool?** Yes, Bot9 enforces usage limits to prevent abuse. If you exceed the allowed requests per minute, subsequent calls will fail until your quota resets. Check the provider documentation for current thresholds. **What should I do if the `create_bot` tool returns an error?** An error usually means a required parameter is missing or invalid. Review the specific error message, which will point to the failing field (e.g., bot name or description). Correcting that single input often solves the issue. **How can I measure the response latency using `send_message`?** You must implement timing logic outside of the MCP call. By recording the time difference between sending the message and receiving the final bot payload, you get accurate performance data. **Can I train my bot with a new URL through the agent?** Yes! Use the `add_data_source` tool and provide the Bot ID and the target URL. The bot will automatically scrape and learn from that page. **How do I review the chat history of a specific user conversation?** First, use `list_conversations` to find the active sessions. Then, use `get_conversation_history` with the Conversation ID to retrieve the full log of messages. **Does the integration allow me to create a completely new bot?** Yes. Use the `create_bot` action and provide a name and a system prompt/instructions. The bot will be instantly created in your account.