# Sonix MCP > Sonix lets you turn raw audio and video files into structured data, summaries, and translations using natural conversation. Transcribe full transcripts in plain text or SRT format, automatically summarize large batches of recordings, and prepare media for global audiences—all without leaving your AI client. ## Overview - **Category:** image-video - **Price:** Free - **Tags:** transcription, subtitles, translation, media-management, ai-summary ## Description Use this MCP to handle all parts of media post-production directly from your agent. Whether you're working with hours of interviews or dozens of podcast clips, you don't have to manually upload them into a separate web portal. You can tell your AI client to transcribe the raw audio and get plain text transcripts instantly. Need subtitles? It handles that too. If you need global reach, it translates those transcripts into multiple languages automatically. Furthermore, if you have a large folder of recordings, you can request a batch summarization for every file at once. This makes your agent act like a true media assistant, handling everything from generating specific video formats to managing user access and organizing the library in folders. ## Tools ### create_batch_summarization Generates a summary report for an entire folder containing multiple media files. ### create_folder Creates a new, organized container (a folder) within your Sonix media library. ### create_media_export Initiates the creation of a downloadable package containing one or more media files. ### create_share Generates a specific link that allows another user to view a selected media file. ### create_summarization Creates a summary report specifically for one single media file. ### create_translation Starts the process of translating an existing media transcript into a different language. ### create_video_burn_in Prepares a video by permanently overlaying subtitles onto the video track, making it ready for social media use. ### delete_media Removes a specific media file from your Sonix account library. ### delete_share Revokes access by removing an existing share link for a given media file. ### get_batch_summarization Retrieves the status and details of a previously requested batch summary job. ### get_media_export Checks the current progress or completion status of a media export request. ### get_media Gets general details and status information for any piece of media in your account. ### get_summarization Retrieves the final summary text or current processing status for a single file's summarization job. ### get_transcript_json Fetches a detailed transcript that includes timestamps linked to specific words spoken in the audio. ### get_transcript_srt Downloads the media's transcript formatted as an industry-standard SRT file, useful for video editors. ### get_transcript_text Retrieves a clean, continuous text dump of the entire audio content without time stamps or formatting. ### get_transcript_vtt Downloads the media's transcript formatted as a VTT file, common for web video players. ### get_translation Checks the status and retrieves the translated content from a previous translation request. ### get_video_burn_in Checks the progress of creating burn-in subtitles on a video file. ### invite_user Invite a new user to the account ### list_folders List all folders ### list_media List media files ### list_shares List shares for a media file ### list_users Lists all user accounts that currently have access to your media library. ### split_transcript Automatically split transcript into subtitles ### submit_media Submit new media for transcription ### update_folder Update a folder ### update_media Update media attributes ### update_transcript Update transcript words and speakers ### update_user Update a user role ## Prompt Examples **Prompt:** ``` List my most recent media files in Sonix. ``` **Response:** ``` I've retrieved your media list. You have 5 recent files, including 'Interview_v1.mp4' (ID: m123) and 'Keynote_Speech.mp3' (ID: m456). Would you like to get the transcript for any of these? ``` **Prompt:** ``` Transcribe this audio file: https://example.com/audio.mp3 in French. ``` **Response:** ``` I've submitted the media for transcription. The Media ID is 'm789' and the status is currently 'processing'. I can notify you once the French transcript is ready. ``` **Prompt:** ``` Generate a summary for media ID m123. ``` **Response:** ``` I've initiated the summarization for 'Interview_v1.mp4'. You can check the result using `get_summarization` with ID 's999' in a few moments. ``` ## Capabilities ### Transcribe audio/video into text The MCP generates transcripts in plain text, SRT, VTT, or JSON formats with precise speaker labels and timestamps. ### Summarize media content You can create a summary for a single file or run batch summarization across an entire folder of recordings. ### Translate transcripts to new languages Automatically process and translate existing text transcripts into dozens of different languages. ### Organize media library structure The agent can list, create, update, or delete folders and individual media files within your Sonix account. ### Prepare video for social sharing Initiate processes to burn subtitles directly onto the video track, creating content ready for immediate upload. ### Manage team access and permissions You can list current users, invite new team members, or generate secure share links for specific media files. ## Use Cases ### Analyzing an Interview Series A journalist has 10 hours of raw interview footage. They ask their agent to list all the files (`list_media`), then run `create_summarization` on each one, and finally compile a master document using the summary reports. This saves days of manual reading. ### Preparing Content for YouTube Launch A content creator finishes an episode. They ask their agent to transcribe it (`get_transcript_text`), then run `create_video_burn_in` so the subtitles are baked into the video, and finally use `create_translation` to get Spanish versions for a dual-market launch. ### Onboarding New Team Members A marketing manager needs to give access to three specific folders of brand assets. They ask their agent to list existing users (`list_users`), then create the necessary folders, and finally use `create_share` to grant temporary viewing rights only. ### Reviewing User Feedback Calls A product manager needs insights from a week's worth of recordings. They ask their agent to process all files in the folder using `create_batch_summarization` and then retrieve detailed, word-level timestamps for key moments using `get_transcript_json`. ## Benefits - Instant transcripts in multiple formats: You get the raw text (using `get_transcript_text`), or structured files like SRT/VTT for direct use in video editing software. No manual conversion needed. - Efficiency through batch processing: Instead of running summarization on 20 clips one by one, you can initiate a whole folder summary using `create_batch_summarization` and check status with `get_batch_summarization`. - Global reach from local files: Need to hit multiple markets? You ask the agent to run a translation via `create_translation`, giving your content instant multilingual visibility for global campaigns. - Full media lifecycle control: From listing all available files (`list_media`) to organizing them into project folders (`create_folder`), you manage the entire asset pipeline without leaving your AI client. - Streamlined collaboration: You can invite team members using `invite_user` and manage access by generating specific, trackable share links with `create_share`, keeping your media organized and secure. ## How It Works The bottom line is you talk to your agent like it’s already connected to your media backend, bypassing manual web portal steps entirely. 1. First, connect your Sonix API key to the Vinkius Catalog using your preferred AI client. 2. Next, give your agent a direct command, like 'Transcribe this video and summarize it,' referencing the media file's ID or location. 3. Finally, the MCP executes the task through its tools, returning status updates and processed data—whether that's plain text, an SRT file, or a summary report. ## Frequently Asked Questions **How do I get a plain text transcript using Sonix MCP?** You request the transcript type 'text' from your agent. This provides a clean, continuous dump of all spoken words without any time stamps or formatting. **Can Sonix MCP summarize multiple videos at once?** Yes, you can use `create_batch_summarization`. You point it to an entire folder, and the MCP handles running the summary job on every file within that container. **What format does Sonix provide for subtitles?** It provides several formats. For professional video editing, you can use `get_transcript_srt` (SRT). For web display, the VTT format is available via `get_transcript_vtt`. **How does Sonix MCP help with team access?** You manage access by listing current users using `list_users`, or you can invite new members and generate secure share links for specific media assets. **Is the translation from Sonix MCP automatic?** Yes, after transcribing a file, you use `create_translation` to automatically process the text into dozens of required languages without human intervention.