Origami
Origami
Documentation

How Origami Works

Everything you need to understand how to fold your documents, repositories, and raw text into interactive AI-powered mini-apps.

Getting Started

Start here

Origami is a universal interactive document engine. Give it any source — a GitHub repository, a PDF, or plain text — and it builds you an interactive, AI-powered workspace in seconds.

1

Choose your source

Select from three intake modes on the home page: Import Repository, Paste Source Text, or Upload Documents.
2

Load the workspace

Origami fetches and parses your source, then drops you into the workspace dashboard with a live source panel, breakdown, and interactive canvas.
3

Generate interactive artifacts

Click Generate Interactive to produce an AI-streamed architecture graph, or v0 MVP to spin up a full single-page mini-app.

Source Types

Origami supports three source types, each with its own intake flow and breakdown experience.

GitHub Repository

Paste any public GitHub URL. Origami fetches every markdown file, manifest, and README across the default branch.

Best for architecture diagrams and dependency graphs

PDF Document

Upload a PDF and Origami extracts all text content page-by-page, then generates a structured insight breakdown.

Best for research papers, reports, and policies

Plain Text

Paste raw text directly or upload .txt, .md, .csv, .json, .ts, or .yaml files for instant AI analysis.

Best for specs, logs, and configuration files

Supported file extensions

.pdf.txt.md.mdx.json.yaml.yml.csv.ts.tsx.js.jsx

Workspace

The workspace is a three-panel dashboard that lets you explore, question, and generate artifacts from your source.

Panel layout

Interactive Canvas

AI-streamed architecture graphs, calculators, and scenario simulators.

Source

Raw content viewer

Breakdown

AI insight cards

v0 MVP

Generated site

Source Panel

View the raw extracted content from your source. For repos, switch between tabs for each file. For PDFs, see the extracted page text.

Breakdown Panel

AI-generated insight cards: summary, key entities, themes, action items, and architecture notes.

Q&A Box

Ask follow-up questions about any source directly inside the workspace. Answers are grounded in the source content.

Repo Navigator

For GitHub sources, navigate between README, CONTRIBUTING, and other docs via the tab bar.

Interactive Canvas

The interactive canvas is the heart of Origami. Clicking Generate Interactive sends your source to the AI, which selects and builds the most appropriate interactive component.

Architecture Graph

repo / text

For repos and technical docs — renders a force-directed node graph of system components and relationships.

PDF Insight Cards

pdf

For PDFs — generates themed breakdown cards with entity recognition, themes, and cross-references.

Scenario Simulator

text

For policy or rule-based docs — simulates real-world scenarios to test outcomes against the source rules.

Package Dashboard

manifest

For manifest files (package.json, pyproject.toml) — renders a dependency and script dashboard.

v0 MVP Generation

AI powered

The v0 MVP feature generates a fully self-contained single-page web app from your source in one click. The output is persisted and accessible via a shareable link.

What gets generated

  • A responsive single-page HTML/CSS/JS app tailored to the source content
  • Styled using a consistent design system derived from the source domain
  • Persisted in browser storage and accessible at /mvp/[id]
  • Copyable as a standalone HTML file with no external dependencies

Flow

Source loaded→Brief generated→POST /api/v0-preview→Artifact persisted→Redirect to /mvp/[id]

API Reference

Origami exposes several internal API routes used by the workspace. These are Next.js Route Handlers under /api.

POST/api/github-repo

Fetches and parses all markdown and manifest files from a public GitHub repository URL.

{ "url": "https://github.com/owner/repo" }

POST/api/pdf-extract

Accepts a PDF file via multipart form and returns extracted text with insight metadata.

FormData with field: file (PDF)

POST/api/document-insight

Runs AI analysis on a text document and returns structured insight cards.

{ "title", "content", "sourceKind", "path", "relatedPaths" }

POST/api/v0-preview

Generates a full single-page MVP site artifact from a source brief.

{ "sourceKind", "sourceLabel", "brief" }

POST/api/source-qa

Answers a user question grounded in the active source content.

{ "sourceKind", "sourceLabel", "brief", "question" }

Ready to try it?

Open the workspace and fold your first document in seconds.

Open Workspace