# Abridge MCP > Abridge handles clinical documentation by taking audio recordings of patient-provider interactions and converting them into structured medical notes. It manages the entire process, letting your AI agent handle everything from starting a new session to retrieving detailed records like the HPI or Assessment & Plan. ## Overview - **Category:** artificial-intelligence - **Price:** Free - **Tags:** clinical-documentation, medical-ai, transcription, healthcare-api, medical-notes ## Description When you connect this MCP, your AI client takes over the messy parts of clinical charting. Instead of spending hours transcribing and formatting notes after an encounter, you simply feed it audio recordings. The process is hands-off: first, the system accepts the audio; second, it processes the conversation in the background; and finally, your agent pulls out perfectly organized documentation sections—including the History of Present Illness (HPI) and the full Assessment & Plan. This turns raw speech into clean, actionable medical records instantly. Vinkius makes accessing this capability easy, letting you connect to Abridge alongside thousands of other tools without complex API juggling. ## Tools ### create_recording Starts a new audio processing session or uploads an existing file containing patient conversation. ### get_clinical_notes Pulls structured data, including the HPI, physical exam findings, and assessment plan from a finished recording. ### get_recording_status Checks if the note generation is pending, processing, or ready by querying the current status of the session. ## Prompt Examples **Prompt:** ``` Create a new clinical recording for patient 'PT-882' using audio file 'encounter_01.mp3'. ``` **Response:** ``` I've started the recording session for patient PT-882. The recording ID is 'rec_abc123'. I will monitor the status for you. ``` **Prompt:** ``` What is the current status of the Abridge recording 'rec_abc123'? ``` **Response:** ``` The recording 'rec_abc123' is currently in the 'processing' stage. It should be completed shortly. ``` **Prompt:** ``` Fetch the final clinical notes and assessment plan for recording 'rec_abc123'. ``` **Response:** ``` I've retrieved the notes. The HPI indicates a 45-year-old male with hypertension. The Assessment & Plan includes starting Lisinopril 10mg daily and a follow-up in 2 weeks. ``` ## Capabilities ### Initiate new audio processing jobs Starts a recording session or uploads an existing audio file for conversion. ### Check job status Determines if the note generation is pending, actively running, or complete. ### Retrieve structured notes Extracts formalized medical documentation, including HPI, physical exam details, and assessment plans, ready for charting. ## Use Cases ### Post-visit Charting A doctor finishes a patient visit. Instead of writing notes by hand, the assistant uploads the audio file using `create_recording`. After waiting for status confirmation with `get_recording_status`, they pull the full documentation via `get_clinical_notes` to draft the chart in minutes. ### Reviewing Historical Encounters A specialist needs context from an old visit. They upload a recorded conversation and ask their agent to pull out only the HPI section using `get_clinical_notes`, skipping over irrelevant administrative chatter. ## Benefits - You eliminate manual transcription. Instead of listening to and typing out every word, the system handles the entire conversion from speech to text automatically. - The `get_clinical_notes` tool provides specific sections like HPI (History of Present Illness) and Assessment & Plan. This structure saves you hours comparing unstructured notes against required chart fields. - You don't have to guess if the notes are ready. Calling `get_recording_status` tells your agent exactly when the process is done, allowing for reliable automation. - It reduces administrative overhead dramatically. Medical scribes can focus on review and refinement instead of starting from a blank page. - The integration supports multiple clients, so whether you use Cursor or Claude, your agent knows how to initiate recording with `create_recording`. ## How It Works The bottom line is that you don't wait around; your AI client manages the sequence of calls until the final document is ready to use. 1. Your agent calls the tool to start a new recording session or upload an audio file. 2. The system processes the conversation. You then use a separate tool call to check the status, waiting until the notes are marked 'completed.' 3. Finally, your agent calls the retrieval tool, pulling structured data like the HPI and Assessment & Plan from the completed recording. ## Frequently Asked Questions **How do I check if my audio file is still being processed?** Use the `get_recording_status` tool with your unique recording ID. It will return whether the task is pending, processing, completed, or failed. **What kind of data is included in the clinical notes?** The `get_clinical_notes` tool returns structured medical data including a summary, HPI (History of Present Illness), physical exam findings, and the assessment and plan. **Can I provide patient and provider IDs when creating a recording?** Yes! The `create_recording` tool accepts optional `patient_id` and `provider_id` strings, as well as a metadata JSON object to provide full context for the encounter. **When using the `create_recording` tool, what authentication method is required for my AI client?** You must provide a valid Abridge API Key in your connection settings. The key authorizes your agent to interact with our service and manage new recording sessions. **If the process fails, how can I use `get_recording_status` to troubleshoot an issue?** The status endpoint will return specific error codes that indicate why processing stalled. These codes help you determine if the failure was due to file corruption or a system timeout. **Can I use `create_recording` with audio files other than standard MP3 format?** Yes, while MP3 is recommended, the API generally supports WAV and FLAC formats. Just ensure your client library handles the file upload correctly for the tool. **Does calling `get_clinical_notes` multiple times quickly trigger any rate limits?** We recommend checking our current usage dashboard for specific rate limits and quotas. High-volume calls might require increasing your API plan to maintain steady access. **What's the recommended workflow sequence when using `create_recording` through to `get_clinical_notes`?** The best flow is: first, use `create_recording`. Second, repeatedly check progress with `get_recording_status`. Finally, once complete, call `get_clinical_notes` to retrieve the structured data.