# Bland AI MCP for AI Agents MCP > Bland AI automates entire phone communication pipelines, letting your AI agents dispatch conversational calls and analyze the results. Connect it to manage inbound numbers, retrieve full call transcripts, or instantly determine if a sales goal was met during a conversation. ## Overview - **Category:** superpower - **Price:** Free - **Tags:** voice-ai, conversational-ai, telephony, automated-calling, webrtc, call-analytics ## Description This MCP gives your AI client total control over enterprise-grade telephony and voice workflows. You can programmatically manage everything from outbound campaigns to complex customer support monitoring. Need to run a large campaign? Use the bulk dispatch tools to send conversational agents to hundreds of numbers simultaneously. Finished a call, but need to know if the agent hit its goal? You pull the transcript and analyze it instantly for specific outcomes. For operations teams, this means you can retrieve full historical call logs or grab raw recordings as MP3 files for quality assurance. It also lets you view all purchased inbound numbers so your AI client knows where to route incoming customer calls. This capability makes Bland AI a central piece of infrastructure that integrates perfectly with the thousands of tools available in the Vinkius catalog, putting telephony control right into your agent's hands. ## Tools ### analyze_call Analyzes an active recording and returns whether a specific business goal was achieved during the call. ### end_call Forcefully disconnects any currently running AI phone call immediately. ### get_batch Retrieves a summary of how many bulk campaigns are running and their current status. ### get_call_details Pulls all specific data points, variables, and the full transcript text for one completed call. ### list_inbound Lists all purchased phone numbers that are available to route incoming customer calls to your AI agent. ### list_calls Retrieves the full, chronological history of every phone call made or received through the system. ### get_recording Pulls a direct link to the raw audio file (MP3/WAV) for deep quality assurance review. ### send_batch Dispatches multiple AI agents at once, scaling up large telecommunications campaigns. ### send_call Sends an automated conversational agent to a single target phone number for immediate outreach. ### create_web_call Spawns a temporary, internet-based signaling socket needed for testing browser audio connections. ## Prompt Examples **Prompt:** ``` I need to run a campaign of 100 calls today targeting enterprise CTOs. ``` **Response:** ``` **Campaign Status:** ✅ Ready to Dispatch. * **Method:** Bulk Batch Send (`send_batch`) * **Targets:** 100 unique numbers (verified) * **Agent Profile:** Sales Outreach v3.2 (Approved) * **Estimated Completion:** Within the next hour. *Note: Keep monitoring the status using `get_batch` for real-time progress.* ``` **Prompt:** ``` What was the outcome of call ID c-1987? Did they want a demo? ``` **Response:** ``` **Analysis Complete.** **Goal Query:** Was the customer interested in a demo? **Result:** TRUE (Success). > The prospect explicitly stated, 'I'm free next Tuesday at 2 PM.' *Recommendation: Schedule follow-up immediately.* ``` **Prompt:** ``` End this active call now. It sounds like the agent is stuck in a loop. ``` **Response:** ``` **Command Executed.** The telephony session for the current call has been forcefully disconnected. A synchronous interrupt flag was sent to Bland AI, and the connection has been successfully shut down. ``` ## Capabilities ### Execute Bulk Campaigns Dispatch single or large batches of conversational agents to target phone numbers at scale. ### Monitor and Control Live Calls Instantly end a live call, analyze recordings for goal completion status, or retrieve detailed transcripts from completed calls. ### Manage Inbound Routing Identify available inbound phone numbers and create web-based signaling sockets for browser audio testing. ## Use Cases ### Evaluating Sales Campaign Outcomes An SDR runs a batch campaign using `send_batch`. After the calls, they ask their agent to use `get_call_details` for five specific IDs. The agent reviews the transcripts and reports back, showing which leads achieved the 'demo interest' goal. ### Handling Critical Support Incidents A support team member notices an agent is repeating bad advice during a live call. They use `end_call` to immediately interrupt and shut down the session, preventing further damage while they step in. ### Building Internal Testing Tools A developer needs to test how their new web application handles voice input. They use `create_web_call` first, then simulate a call flow using the resulting signaling socket, all within their IDE. ### Auditing Recorded Interactions The QA manager needs to check audio quality for 50 calls. Instead of relying on text summaries, they use `get_recording` for each call ID, pulling the native MP3 files directly into a bulk analysis folder. ## Benefits - Determine lead qualification status instantly. Use the `analyze_call` tool to query transcripts and confirm if a specific goal was met without manual review. - Scale your outreach campaigns immediately. The `send_batch` tool lets you dispatch hundreds of agents concurrently, far faster than any manual dialing system. - Keep full control over live interactions. If an agent gets stuck or goes off-script, the `end_call` tool lets you shut down the session instantly. - Streamline data collection. The `get_recording` tool pulls native MP3/WAV files directly, so you never have to rely on a limited text transcript for QA logging. - Manage your entire phone presence. Use `list_inbound` to see all available numbers and `send_call` to test agent performance against a single target number. ## How It Works The bottom line is, once connected, you treat the entire telecom system like another toolset within your AI client's natural conversation flow. 1. Connect your Bland AI API key to Vinkius, then link the MCP to your preferred AI client (Claude, Cursor, etc.). 2. Use your agent to list available inbound phone numbers or retrieve historical call logs. 3. Issue a command to dispatch agents, analyze a transcript, or end an active call. ## Frequently Asked Questions **How do I run bulk campaigns using Bland AI with my AI agents?** You use a batch dispatch tool. This lets your agent send out automated calls to dozens or hundreds of numbers at once, scaling up outreach quickly without manual intervention. **Can I analyze recordings from past calls using Bland AI with my agents?** Yes, you can. You pull the transcript and then ask your agent to run a targeted analysis on it, determining if specific goals—like scheduling or interest level—were met. **What is the best way to monitor incoming calls with Bland AI?** First, you use the list inbound tool to check your purchased numbers. Then, you connect these numbers so they can be routed directly to your automated agent for handling. **If an active call goes wrong, how do I stop it using Bland AI?** You issue a command to end the call. This sends an immediate interrupt signal to the phone system and shuts down the session instantly, letting you take over. **Does Bland AI help me debug my code when testing web audio?** Absolutely. You can use a WebRTC signaling socket creation tool. This lets developers test browser-based audio flows directly within their coding environment for accurate debugging.