Raggamuffin | Merge • Ask • Listen

Raggamuffin turns static PDFs into an interactive experience. Merge documents, ask questions, and listen to clear, grounded answers: fast, lightweight, and designed for real teams.

Comment

Inspiration

Raggamuffin was born from a simple gap we kept seeing in PDF workflows: people have piles of documents and no easy, fast way to merge them, extract knowledge, and interact with them in real time. Most tools either do basic merge operations or offer AI chat in locked-down platforms with high cost and low transparency.

We wanted a clear, developer-friendly system that turns PDFs into a living interface: merge, ask questions, and hear answers—all without vendor lock-in or heavy infrastructure.

What it does

Raggamuffin is a modular, API-first platform for PDF merging and document chat.

At its core, it provides:

  • PDF merging and download
  • RAG-style question answering over uploaded PDFs using Gemini
  • Optional voice playback via ElevenLabs
  • A lightweight frontend that can be hosted on GitHub Pages
  • A clean backend API deployable to Cloud Run
  • Use cases include document review, research, education, and team knowledge extraction.

How we built it

The system is built as a simple, composable stack with a focus on reliability and fast demo readiness.

Key design choices:

  • FastAPI backend with clear endpoints for merge, chat, and audio
  • Gemini-powered retrieval pipeline (chunking + embeddings + context generation)
  • Static frontend designed for fast deployment via GitHub Pages
  • Cloud Run deployment for low-ops, scalable hosting
  • Optional integrations (ElevenLabs, Datadog, Confluence, Drive)
  • The project is designed to be usable locally and deployable in minutes.

Challenges we ran into

Balancing simplicity with functionality was the main challenge. We wanted a clean demo without sacrificing core features.

Other challenges included:

  • Handling ephemeral storage on Cloud Run
  • Avoiding UI complexity while keeping the experience polished
  • Managing CORS and multi-origin deployment (Cloud Run + GitHub Pages)
  • Ensuring the system remains stable under free-tier constraints

Accomplishments that we're proud of

  • A working end-to-end PDF merge + chat system
  • Clean, modular backend with optional integrations
  • Fast, low-cost deployment with Cloud Run + GitHub Pages
  • A demo-ready interface with minimal setup friction
  • A transparent pipeline that can be extended without re-architecture

What we learned

Document AI is less about model hype and more about orchestration and usability.

We learned that:

Clarity and speed of deployment matter more than feature breadth Free-tier constraints force good architectural discipline Deterministic pipelines build trust in AI systems A simple UI is often the most effective for demo success What’s next for Raggamuffin Next steps focus on hardening and scale:

Persisted storage (Cloud Storage or DB) for longer-lived PDFs Stronger mobile UI optimization Multi-provider LLM support Better observability around latency and usage Optional premium features (team workspaces, sharing, access control)

Built With

Share this project:

Updates