Agents In Use
What’s new todayLast update
1005 additions

Use Replit Agent to build a document chatbot

Justin Lahullier spent two weeks directing Replit Agent to build a personal chatbot with document-processing features.

Published linkedin.comAgent: Replit Agent
OUR ADAPTATION

Suggested prompt

Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.

Build a personal document chatbot with [file types], saved conversations and selectable [model providers]. Document its data flow and return previews and limitations for my review.

What happened

Lahullier reports a functioning personal document chatbot integrating OpenAI, Anthropic and Google models after two weeks of directing Replit Agent.

Reported by Justin Lahullier

How the agent did it

  1. Describe the document chatbot.
  2. Implement model providers and data storage.
  3. Review the interface and documentation.

Where a person stepped in

Lahullier directed the agent over two weeks.

Justin Lahullier spent two weeks directing Replit Agent to build a personal chatbot with document-processing features. He supplied natural-language instructions while the agent handled architecture and implementation. The resulting application combined multiple AI providers rather than depending on one model.

The account names React, TypeScript, WebSockets and PostgreSQL among the components. It describes real-time communication, document analysis, a responsive interface and integrations for OpenAI, Anthropic and Google models. Lahullier says the agent also created tests, security evaluations and documentation alongside the application. These are descriptions of his development result, not checks independently repeated by this directory.

Human involvement remained substantial: the project took two weeks of his time giving direction, despite his statement that he wrote no application code himself. He reports spending just over one hundred dollars. The article includes further architecture and feature details for the platform. It documents commissioning a custom document tool while leaving claims about reliability and production readiness attributed to its builder.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: www.linkedin.com · Justin Lahullier

Frequently asked questions

Which providers were integrated?

OpenAI, Anthropic and Google.

What database was named?

The application used PostgreSQL.

Who directed the development?

Lahullier directed the agent for two weeks.

More details

About this account

Replit handled architecture and implementation.

Firsthand account opened and read. Results are attributed to the author. Prompt is our suggested adaptation, not an original quotation.

Credit: Justin Lahullier

Source published 29 Apr 2025

Source checked 14 Oct 2026

Tools and topics

Replit Agent · React · TypeScript · PostgreSQL · OpenAI · Claude · Gemini

Document analysis · Multi-model chat · App building

Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.