Skip to content

arnxv0/Pitchweaver

Repository files navigation

PitchWeaver

PitchWeaver is an AI-powered B2B partnership pitch generator. You enter a target company website and your product details, and it produces a polished, ready-to-send PDF pitch deck plus a separate password-protected internal strategy PDF — complete with charts and tailored integration scenarios.

Demo: https://youtu.be/_wYv3TAT6fo
GitHub: https://github.com/arnxv0/Pitchweaver
Figma prototype: https://www.figma.com/design/VTOLvNRDeVdZgCzIpgUdwz/PitchWeaver-Multi-Page-Web-App


The Problem

Partnership and sales teams waste hours on the same repetitive work before every outreach: researching what a company does, who they sell to, what they recently launched, and how to position an integration. That research then has to be written up into a pitch deck, formatted, revised, and packaged — every single time, for every single company.

The documents themselves never stay current either. As companies update their positioning, pricing, and products, the pitch materials become stale. Teams end up maintaining multiple versions of the same deck, manually updating slides, or just sending generic decks that do not feel tailored at all.

PitchWeaver eliminates the research-to-document cycle entirely. It handles web research, writing, data visualization, and professional PDF generation in one automated pipeline.


What it generates

For each pitch you get two documents:

External pitch deck (PDF)

  • Executive summary written to the specific company
  • Target company overview pulled from real-time research
  • Integration scenarios and value proposition mapped to their context
  • Four charts: market growth, integration impact, competitive positioning, roadmap
  • Call to action

Internal strategy notes (PDF, password-protected)

  • Risks and objections
  • Ideal contact roles
  • Conversation starters
  • Internal next steps

How it works

  1. You.com pulls the target company's website content, runs targeted web search queries, and fetches recent news
  2. Gemini Flash converts the research bundle and your product details into structured pitch content (scenarios, ROI bullets, chart data)
  3. Plotly generates four high-resolution chart images
  4. Foxit Document Generation API fills the DOCX templates with content and charts and produces both PDFs
  5. Foxit PDF Services API applies password protection to the internal notes PDF

Sponsors

Foxit Software: Document Automation Hackathon

Foxit wants to see both the Document Generation and PDF Services APIs used in a meaningful end-to-end workflow, not just isolated API calls. The judging focuses on workflow design, real-world usefulness, technical execution, and clear setup.

PitchWeaver uses both APIs in sequence as the final output stage of the pipeline:

Document Generation API generates two PDFs from structured DOCX templates — one external pitch deck and one internal notes document. The templates use Foxit's placeholder syntax for text fields, loops over lists (integration scenarios, ROI bullets), and image injection for the four charts. This is the right tool for this step because the output needs to be a consistently formatted, professional PDF, not just exported text.

PDF Services API applies password protection to the internal notes PDF. The internal document contains sensitive strategy and should not be forwarded externally, so protection is applied programmatically rather than left to the user.

The complete Foxit flow: generate charts → inject into templates → Document Generation API produces PDFs → PDF Services API secures the internal one → both returned to the user.


You.com: Build Intelligent Applications with You.com APIs

You.com wants projects that use their APIs to build applications capable of reasoning and searching live — combining real-time web data, structured research, and citation-backed results to produce reliable, trustworthy outputs.

PitchWeaver integrates all three of You.com's research APIs:

Contents API is called with the target company URL and returns a clean Markdown version of the page. This works even on JavaScript-rendered sites and requires no HTML parsing, which makes it reliable across any company website.

Unified Search API is called with three targeted queries per company — one for product and use case overview, one for integrations and developer ecosystem, and one for pricing and target market. This surfaces context that the homepage alone does not cover.

Live News API is called to pull the five most recent relevant articles about the company. Referencing a recent funding round, product launch, or partnership in the pitch is what makes it feel researched rather than templated.

All three outputs are bundled into a structured research context and passed directly to Gemini as grounded input.


Kilo: Finally Ship It

Kilo wants to see something creative, well-executed, and actually shipped — judging on whether the idea is worth building and whether it was built well.

PitchWeaver has a lot of moving parts for a solo hackathon project: a full React frontend, a FastAPI pipeline with SSE streaming, four third-party API integrations, chart generation, DOCX template manipulation, and two-format PDF output. Kilo Code was used throughout to scaffold the structure, write and debug service modules, build React components, and refactor as the pipeline evolved. It made it possible to build everything to a finished standard rather than cutting corners on error handling, code quality, or the frontend experience to hit the deadline.


Figma: From Idea to Interactive

Figma wants a clickable prototype started with Figma Make covering 3–5 connected screens with one clear user flow, plus a clear explanation of the problem and how the idea could grow.

Before writing any code, Figma Make was used to prototype the full user flow in plain language: landing page with the value prop, two-step pitch creation form, live generation progress tracker, and results page with dual PDF preview and downloads. That became the blueprint for the React frontend so the flow never had to be figured out mid-build. The prototype is linked above.

The problem is clearly defined (research and document prep overhead in B2B sales and partnerships), the target user is sales and partnerships teams at B2B companies, and the idea can grow into a full self-serve SaaS with saved templates, CRM integrations, and team collaboration.


Tech stack

Layer Technology
Frontend React 18, TypeScript, Vite, Tailwind CSS
UI Radix UI, MUI, Recharts, React Router
Backend FastAPI, Python 3.11, SQLite (aiosqlite), Pydantic
Research You.com (Contents, Search, News APIs)
LLM Google Gemini Flash
Charts Plotly + kaleido
PDF gen Foxit Document Generation API
PDF process Foxit PDF Services API

Setup and run

Requirements

  • Python 3.11+
  • Node.js 18+
  • API keys for You.com, Google Gemini, and Foxit

1. Clone

git clone https://github.com/arnxv0/Pitchweaver
cd Pitchweaver

2. Backend

cd pitchweaver-backend
pip install -r requirements.txt

Create .env inside pitchweaver-backend/:

YOU_API_KEY=your_you_api_key
GEMINI_API_KEY=your_gemini_api_key
FOXIT_CLIENT_ID=your_foxit_client_id
FOXIT_CLIENT_SECRET=your_foxit_client_secret

Start the backend:

uvicorn main:app --reload --port 8000

3. Frontend

cd ..
npm install
npm run dev

Open http://localhost:5173. Backend runs at http://localhost:8000.


About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors