Event content
About the webinar
MCP (Model Context Protocol) is an open standard for how AI applications discover and call external tools. Point an MCP client at a server and it reads the tool list and input schemas, then calls those tools on your behalf. The Sarvam Voice Agents MCP server puts the whole platform behind that interface, including agents, telephony, campaigns, evals, knowledge bases and analytics.
A voice agent keeps changing after launch. You tune the prompt, add knowledge, rerun evals and dig into failed calls, each on a separate screen. Over MCP, your coding tool runs that loop from one conversation. Ask it to tighten the prompt, rerun the eval suite and commit if it passes, and it runs each step in order. Edits go to a draft, and only versions you commit take live calls.
What you’ll see
- Connecting a coding tool to the server and signing in through Sarvam SSO
- Creating an outbound agent from a prompt, adding a knowledge base and committing a version
- Running an eval suite against that version and reading the failures
- Placing a test call to a real phone number
- Creating a campaign and triggering real calls to people’s phone numbers
- Pulling last week’s failed calls, grouped by reason
Who should attend
- Developers who want to build voice agents without leaving their coding environment
- Product and engineering teams evaluating voice agents for production
- Current Sarvam Voice Agents users who want to manage agents, evals and campaigns from their editor
Format: Live demo and Q&A
Registration
Add to calendar
Go from idea to production Start shipping today.
Go from idea to production
Start shipping today.