Impulse Propal Generator - Paris Vibe-a-thon: “Build a One-Person Unicorn” — November 15, 2025
AI Tinkerers - Paris
Hackathon Showcase

Impulse Propal Generator

Team consisting of Impulse Lab CEO Leonard (ESILV), CTO Michel, and Software Engineer Mattéo — full‑stack JavaScript/TypeScript (React, Node.js, Docker), AI agents, real‑time systems.

3 members Watch Demo

We are a team from https://impulselab.ai

We wanted to build a propal generator, when we are working on deals for client, it takes a lot of time to go from meeting to the proposal and we want to have a quicker way to do that

We had the idea to use an AI agent (a little bit like Cursor) but for client proposals, so our team could delegate this work to the AI agent and free up time.

Partner Tech Usage:
ElevenLabs: Speech-to-text API for voice input
Qdrant: Vector database for client information retrieval
Google Gemini: AI model with thinking capabilities for agent reasoning

Technologies Used
Frameworks & Libraries: Next.js 16, React 19, TypeScript, Tailwind CSS, TipTap (rich text editor), oRPC (type-safe RPC), Drizzle ORM, Zod (schema validation), React Hook Form, TanStack Query
AI & ML: Google Gemini 2.5 Pro (AI SDK), ElevenLabs Speech-to-Text API, Qdrant Vector Database
Infrastructure: PostgreSQL (via Drizzle), Vercel (hosting), Vercel Blob Storage (image uploads)
APIs & Services: ElevenLabs STT API, Qdrant REST API, Google AI SDK

We used a component library from AlignUI https://www.alignui.com/

Apple ElevenLabs Figma Google Qdrant Vercel n8n

Production deployment

Summarizing URL...

Github repo

Summarizing URL...

Our website

Summarizing URL...