The company behind the thesis.

Why this problem?

Fluent language alone does not make an agent useful. An agent needs to understand the context of a request, recognize when to ask a person, and leave a record of its actions. MershLab studies how those requirements fit together, from human communication to the infrastructure that supports it.

Read the lab's thesis and its research into intent, tone, and human communication.

Saturday is the first public build.

Saturday is our agent harness: a place to connect models, tools, approvals, and checkpoints across interfaces. It makes the execution layer concrete while the broader research continues.

How the work connects

• MershTrust explores verifiable records of AI execution.
• MershBrain brings company knowledge into agent workflows.
• AutoMarket applies agent workflows to business conversations.
• Dispatch is a prototype for auditable voice intake.

These projects are at different stages of development. The work pages describe their current scope; the research notes explain the questions behind them. The broader human intelligence vision remains an active research direction.

Who is building MershLab?

Mohak Gupta is the founder and CEO. Nishant Luhera is the CTO. Meet the team and explore the technical work linked from each project.

Start a conversation.

For investment conversations, technical diligence, or an introduction to the lab, email hello@mershlab.com. Tell us what you would like to explore so we can bring the relevant work to the conversation.