Deepgram MCP. Run STT, TTS, and full API management via chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Deepgram MCP Server. Run full audio AI workflows from your agent. This server handles high-speed speech-to-text (STT) and text-to-speech (TTS) tasks, letting you process audio streams, generate voices, and manage the underlying API infrastructure—all via natural language commands.
You can transcribe remote audio URLs, generate audio from text, and audit usage and keys without leaving your development environment.
What your AI agents can do
Create key
Generates a new, specific Deepgram API key for your account.
Delete key
Removes an existing Deepgram API key to mitigate potential security risks.
Get balances
Retrieves detailed logs showing your current cloud billing limits and thresholds.
You pass a web URL, and the agent processes the audio stream using the Nova-2 model to return a text transcript.
You provide raw text and a voice selection, and the agent generates and returns the resulting audio binary stream.
The agent queries usage data, returning specific metrics on transcription time and TTS byte usage.
You can list, create, or delete Deepgram API keys for different projects.
The agent identifies and lists all associated Deepgram projects and members.
You retrieve detailed cloud logging tracing your account's vault limits and current usage against thresholds.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
019d7583create key
Generates a new, specific Deepgram API key for your account.
019d7583delete key
Removes an existing Deepgram API key to mitigate potential security risks.
019d7583get balances
Retrieves detailed logs showing your current cloud billing limits and thresholds.
019d7583get usage
Calculates and reports the properties driving your active account usage metrics.
019d7583list keys
Lists all existing Deepgram API keys associated with your account.
019d7583list members
Retrieves a list of all members and users within a specific Deepgram project.
019d7583list projects
Identifies and lists all distinct Deepgram projects you have set up.
019d7583send invite
Sends a team invitation to a user for a specific Deepgram project.
019d7583speak text
Generates an audio binary stream from text input using specified voices and billing rules.
019d7583transcribe url
Processes a web URL containing audio and returns the full text transcription using the Nova-2 model.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Deepgram, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
You can run your full audio AI workflows straight from your agent using this Deepgram server. This thing handles high-speed speech-to-text (STT) and text-to-speech (TTS) tasks, letting you process audio streams, generate voices, and manage the underlying API infrastructure all through natural language commands. You'll never have to leave your dev environment.
To get a transcript from an audio file online, you just pass a web URL to transcribe_url, and the agent processes that stream using the Nova-2 model to return the full text transcript.
Need to generate audio? Give the agent raw text and a voice selection to speak_text, and it spits out the resulting audio binary stream following your billing rules.
For managing your access, you can list_keys to see every API key you've got, create_key to generate a new, specific key, or delete_key to pull an old one to keep things secure.
If you're dealing with multiple groups or projects, you'll use list_projects to see all your Deepgram projects, and you can list_members to get a roster of every user in a specific project, letting you send_invite to bring new people into the fold.
Checking the books is simple. You can get_usage to calculate and report the specific metrics that drive your active account usage, and you'll use get_balances to pull detailed logs that trace your current cloud billing limits and thresholds.
To make sure your pipelines don't drop due to access issues, you can also list_members and list_projects to organize your tenants. You'll also find that get_usage gives you the properties driving your account usage metrics, and get_balances shows you exactly what your vault limits are.
How Deepgram MCP Works
- 1 First, subscribe to the Deepgram server and provide your Deepgram API Key (it's in the Deepgram Console).
- 2 Next, you tell your agent what you need—for example, 'Transcribe this URL' or 'Generate speech for this text'.
- 3 The agent executes the necessary tool call, and the results (transcripts, audio streams, or usage reports) appear directly in the chat.
The bottom line is, you manage complex audio AI workflows and API credentials using natural conversation.
Who Is Deepgram MCP For?
This is for developers and data engineers who need to test STT/TTS models and audit audio pipelines without leaving their development environment. If your job involves analyzing voice data, managing API costs, or integrating third-party audio services, this is your toolset.
Tests STT/TTS models and manages API keys directly within the chat interface, accelerating the development cycle.
Audits transcription volumes and manages project-wide audio pipelines by issuing commands like 'Show usage for Project X'.
Monitors audio AI usage in real-time, verifying transcription accuracy and tracking costs to hit product goals.
Tracks wallet balances and manages team access across multiple Deepgram projects to keep the entire system operational.
What Changes When You Connect
- Transcribe audio from remote URLs. Use
transcribe_urlto process any WAV/MP3 web stream, getting text transcripts instantly via the Nova-2 model. - Generate synthetic voices. Use
speak_textto create high-fidelity audio from raw text, letting you test TTS outputs before deployment. - Manage billing from chat. Check your account health with
get_balancesandget_usageto see your exact API consumption and vault limits. - Control who has access. List projects using
list_projectsand invite new team members withsend_invite—all without leaving your agent. - Improve auditability. Use
list_keysandcreate_keyto track and manage every API key, keeping your security boundaries tight. - Simplify workflows. Your agent handles the complex data calls, allowing you to audit data and manage resources with simple commands.
Real-World Use Cases
Debugging a Transcription Pipeline
A data engineer needs to verify if a new transcription service is working. They run transcribe_url on a test recording, get the text output, and then use get_usage to see exactly how much transcription time the test consumed. This confirms both functionality and cost.
Onboarding a New Team Member
A product manager adds a new team member to the project. They run list_members to verify existing users, then use send_invite to grant access to the new user, and finally use create_key to issue their dedicated API credentials.
Creating Automated Voice Content
A content creator needs an audio file for a blog post. They use speak_text with a specific voice and text. The agent returns the binary audio stream, which they can immediately download and use.
Checking API Limits Before Launch
A DevOps engineer prepares for a major launch. They run get_balances to confirm the current vault limits are above the projected peak load, and they run list_projects to ensure all necessary environments are accounted for.
The Tradeoffs
Using the console GUI for everything
Manually navigating Deepgram's console to list keys, check usage, and generate a new key takes multiple logins, copy/pastes, and context switches. It's slow and hard to script.
→
Use the Deepgram MCP Server. Your agent runs list_keys to see all keys and create_key to generate a replacement. You manage all keys and usage metrics from one conversation.
Ignoring project boundaries
Trying to process an audio file without knowing which project's credentials to use, leading to billing confusion or using stale API keys.
→
First, use list_projects to identify the correct project UUID. Then, manage access using list_members and send_invite to ensure the right people have the right keys.
Overlooking usage costs
Running several long test transcribations without checking the API cost first, only realizing the high usage cost when the billing dashboard shows an unexpected spike.
→
Always run get_usage before heavy testing. This tool reports your current transcription time and TTS byte usage, letting you budget your tests.
When It Fits, When It Doesn't
Use this server if your workflow requires constant interaction with Deepgram's core features: STT, TTS, and the underlying account management. You need to perform actions like checking limits (get_balances), creating keys (create_key), or managing users (list_members) alongside running audio tasks. Don't use it if you just need to run a single, isolated transcription job. For that, a simple API wrapper might suffice. This server is for orchestration and auditing. If your primary need is just generating a single audio file, you can use speak_text alone. But if you need to track the cost of that file, you need the full set of tools.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Deepgram. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Handling audio data usually means jumping between three different dashboards.
Today, processing audio data means jumping between the Deepgram console for key management, the billing portal for usage limits, and a separate transcription page for the actual audio job. You copy the URL from one tab, paste it into another, and then copy the resulting text to a third system. It's a lot of context switching.
With this MCP server, you tell your agent, 'Transcribe this audio URL.' The agent runs `transcribe_url`, handles the connection, and gives you the text result immediately. You can then ask, 'How much did that cost?' and the agent runs `get_usage` right there.
Deepgram MCP Server: Run full key and audio management.
You no longer have to manually list keys, check project boundaries, and send invites through separate UIs. Your agent handles the sequence. You ask to 'list all projects and add a user,' and the agent runs `list_projects`, followed by `list_members` and `send_invite` automatically.
It brings all the core functionality—from key creation to user management—into one conversation. You get a single, unified control plane for your entire Deepgram stack.
Common Questions About Deepgram MCP
How do I transcribe an audio file using the Deepgram MCP Server? +
You run the transcribe_url tool. You just provide the full web URL of the audio, and the agent returns the text transcript. It uses the Nova-2 model for the job.
Can I generate speech using the Deepgram MCP Server? +
Yes, use the speak_text tool. You give the text and the desired voice, and the agent returns the high-fidelity audio binary stream.
What is the best way to check my Deepgram usage with the Deepgram MCP Server? +
Use the get_usage tool. It reports specific metrics on your transcription time and TTS byte usage, which is better than just looking at a total dollar amount.
How do I create a new API key using the Deepgram MCP Server? +
Call the create_key tool. The agent handles the process and provides the new key string, which you then need to secure.
What does `list_projects` do in the Deepgram MCP Server? +
list_projects identifies and gives you a list of all Deepgram projects you have set up. This helps you know which project you are currently working in.
How do I use `get_balances` to check my Deepgram wallet limits? +
get_balances retrieves explicit cloud logging tracing explicit Vault limits. This tool confirms your current billing thresholds, ensuring your audio pipelines don't stop unexpectedly.
What does `list_members` do when I need to manage my Deepgram team access? +
list_members dispatches an automated validation check routing explicit Gateway history. You use this to see who has access and manage your developer team's permissions.
How do I create an API key for a specific Deepgram project using `create_key`? +
create_key identifies precise active arrays spanning native Gateway auth. Running this tool generates a new, isolated API key for a specific project, keeping your credentials secure.
Can my agent transcribe an audio file from a public URL? +
Yes. Use the 'transcribe_url' tool. Provide the public URL of the audio file (WAV, MP3, etc.) and specify the model (e.g., 'nova-2'). The agent will dispatch the request to Deepgram and return the transcribed text instantly.
How do I generate speech from text using the agent? +
Use the 'speak_text' tool. Provide the text script and the target voice model (e.g., 'aura-asteria-en'). Your agent will trigger the high-fidelity Aura voice engine and return the binary audio stream data.
Can I monitor my remaining project balance via chat? +
Absolutely. Use the 'get_balances' tool with your project ID. The agent will retrieve your current wallet thresholds and funding limits directly from Deepgram to ensure your audio pipelines stay active.
Multi-server workflows that include Deepgram MCP
Create AI Podcast Content Using MCP Servers
You record a 45-minute podcast, spend 4 hours editing the transcript, and still do not have show notes, a blog post, or social clips , because transcription tools give you text but not intelligence
MCP Servers for Automated Visual Podcasts
Your podcast is audio-only and invisible on YouTube and Instagram , Leonardo AI generates cinematic visuals for every segment, Deepgram timestamps every word for sync, and Sheets manages the production pipeline
MCP Servers That Remember Every Meeting
You had a critical decision in a meeting 3 weeks ago but nobody remembers the exact reasoning , Deepgram transcribes every meeting, Mem0 stores decisions with persistent memory, and Sheets tracks all commitments
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Scale AI
Automate high-quality data labeling and RLHF workflows via Scale AI — create projects, manage batches, and submit annotation tasks directly from your AI agent.
Exa AI
Search the web with neural embeddings that understand meaning, not just keywords, and return the most relevant results for any query.
Retell AI
Empower your conversational AI to orchestrate, analyze, and automate phone calls or web-based voice agent interactions via Retell.
You might also like
Eventbrite
Create events, sell tickets, and manage attendees with the world largest self-service ticketing platform for any occasion.
Unbounce
Automate marketing tasks via Unbounce — retrieve landing pages, fetch captured leads, audit performance stats, and manage test variants easily.
Metaplane
Data observability via Metaplane — track monitors, incidents, and data quality metrics.