StockPile

Inspiration

Nikhil and Antony were grabbing coffee and Nikhil asked him about problems he knew about. Antony is COO at a $10MM ARR ecommerce startup and chatted about a recent challenge balancing inventory and production planning. Nikhil figured he could build an agent to tackle that.

What it does

The agent imports information regarding current inventory levels by warehouse, a retail and DTC forecast and cost constraints. It then uses code-interpreter to read the data and develop a shipping, logistics and production plan. It automatically debugs it's own code using toolhouse and checks the output matches the constraints before presenting the results to the user. The next step would be to build additional agents which can execute the plan.

How we built it

Python, Cursor and Toolhouse.

Challenges we ran into

Toolhouse wasn't correctly reading the error codes back to the agent, the founder figured out the fix and pushed it live on the spot - way to go!!!

Accomplishments that we're proud of

Building a working agent that solves a real business problem in just a few hours.

What we learned

Agents are the future and it's already here

What's next for StockPile

This agent is just the first step in what could be a whole network of agents that could automatically manage operations and supply chain for an ecommerce company. These agents have the potential to decentralize a companies operations, simplifying and increasing team efficiency. YC actually has 2 requests for startups addressing these types of challenges, agents might be the answer to competing with the systems of record (Oracle, Netsuite, SAP etc) that run companies. We're going to explore the feasibility of using agents for this.

Built With

• claude
• python
• toolhouse