GrillMaster Recruiter

• 

  Next page ->

• 

  Le UI

One-Liner:

Next-gen recruiting pipeline with resume bullsh*t prescreening using real-time voice agents and MCP tool-calling to grill candidates over phone.

Inspiration

In the Bay Area alone, thousands of candidates apply to IT positions daily. But with such high volume, recruiters face a critical dilemma: how to quickly identify top talent without overlooking hidden gems? Traditional recruiting pipelines can't scale — and exceptional candidates slip through the cracks. GrillMaster was born to solve this challenge using Agentic AI, MCP (Multi-Channel Processing) tool-calling, and real-time voice agents: offering a next-gen, scalable pre-screening solution that lifts the load off recruiters and helps the best candidates shine through.

What it does

GrillMaster provides a convenient pre-screening process that is responseive to users and provides additional info to recruitiers. users can submit their linkedin profile, email and phone number to be connected to a voice agent that will interview them to see how fit they are for the position, using MCP calls to process their linkedin profile. When the agent is done, it will email its referral to the recruiter and user through mcp. Afterwards, the recruiter can observe the AI summary of the conversation and referral opinion to decide whether they should proceed with the hiring process or not.

How we built it

We built GrillMaster with Twilio webhooks running on our FastAPI app. For the AI , we used OpenAI Realtime API combined with OpenAI's MCP server to allow connection to LinkedIn and Email. For the frontend we created a simple Nextjs app.

Challenges we ran into

One challenge we encountered was injecting MCP connections into OpenAI's realtime API, since this was not supported by default. to fix this, we wrote a connection layer that would parse the MCP functions and inject them using OpenAI's function calling system, as well as a custom event handler to call the MCP servers.

Accomplishments that we're proud of

We are proud of assembling such strong and useful MCP server together with an advanced voice agent. We think this combination allows for truly powerful next-generation levels of ai intelligence and efficiency.

What we learned

We learned how to provide the convenience of MCP -- insane strong tools at your fingertips -- to a voice AI agent, allowing for seamless audio interaction through phone.

What's next for GrillMaster Recruiter

Aside from applying to YC: We hope to improve the app further to allow more advanced features including agentic AI research, as well as attaching more MCP servers like GitHub and Notion. We were also close to implementing a system that would allow the AI to schedule a follow-up in-person interview through MCP based on oberving the google calendar schedule of both the recruiter and user at a time that is convenient for both, and we would like to finish that.

Built With

• agents
• emailmcp
• googlecalendarmcp
• linkedinmcp
• mcp
• openai
• python
• realtimespeechtospeech
• voiceagent