Share knowledge in one link.

Context sharing for AI and humans.
Paste knowledge → get a short link. Humans see a rich page. AI agents see clean structured data. Same URL, two rendering paths.

Quick Start • How It Works • API • Stack • Contributing

---

The Problem

Knowledge lives in Slack threads, meeting transcripts, Figma files, and RFCs. The hardest part of every conversation is getting everyone up to speed. And when someone finally shares context, the recipient has to re-format it for whatever tool they use next.

AI agents can't read your Notion doc. ChatGPT doesn't see your Slack history. Gemini doesn't know what you discussed in that meeting.

Context gets lost. Conversations start from zero. Every time.

---

The Solution

Contxt is a single link that serves both humans and AI.

Paste your knowledge — analysis, RFC, meeting recap, research brief — and get a short URL. Share it anywhere. When someone opens it:

• Humans see a rich card with title, summary, full markdown, and buttons to continue in ChatGPT, Gemini, or Claude.
• AI agents (ChatGPT, Gemini, Claude, curl) receive structured YAML — clean, no HTML/CSS noise, ready to continue the conversation with full context pre-loaded.

One link. Two rendering paths. Zero friction.

                ┌─ User-Agent: browser ──► Rich HTML page Continue in ChatGPT / Gemini / Claude
                │
  contxt.to/s/abc ──┤
                │
                └─ User-Agent: AI agent ──► Structured YAML (text/plain) name, description, content

---

Features

• Paste and share — Enter content, get a short link. No account required.
• AI-native — Same link works in ChatGPT, Gemini, Claude, and any HTTP client. AI agents get machine-readable YAML automatically.
• Private by default — Nothing is public without you sharing the link. No indexing, no discovery.
• Dual-render architecture — Server-side user-agent detection serves the right format every time.
• API-first — Create contexts programmatically via POST /api/contexts.
• Self-hostable — Open source. MIT license. Your data, your infrastructure.

---

Quick Start

git clone https://github.com/redirhub/contxt.to.git
cd contxt.to
cp .env.example .env.local
npm install
npm run dev

Open http://localhost:3000.

Configuration

Create a .env.local file based on .env.example:

Variable	Required	Description
TURSO_DB_URL	Yes	Turso (libSQL) database URL
TURSO_DB_AUTH_TOKEN	Yes	Turso authentication token
AUTH_GOOGLE_ID	No	Google OAuth client ID
AUTH_GOOGLE_SECRET	No	Google OAuth client secret
AUTH_SECRET	No	NextAuth encryption secret
EMAIL_SERVER_URL	No	SMTP URL for magic link emails

Running Tests

npm test               # Unit tests (Vitest)
npm run test:e2e       # E2E tests (Playwright)
npx tsc --noEmit       # Type check
npm run lint           # Lint

---

How It Works

Creating a Link

1. Go to contxt.to — or use the API.
2. Paste or write your content. Optionally provide a title.
3. Click Create — AI generates a title and summary from your content.
4. You get a short link: contxt.to/s/quick-thought-kF3mQ.

Reading a Link

As a human: Open the link in any browser. You see a card with the title, summary, full markdown content, and action buttons to continue the conversation in any major AI.

As an AI agent: Visit the same link. The server detects your user-agent and returns structured YAML via Content-Type: text/plain — no HTML, no CSS, just the data:

name: Q1 Design Sprint Retro
description: Summary of 3 design sprints — onboarding hit 87% usability, analytics needs iteration.
content: |
  We ran 3 sprints this quarter...

  Key wins: the onboarding redesign tested at 87% usability (up from 62%).
  Key misses: the analytics dashboard overhaul needs another iteration.

API

Create a context programmatically:

curl -X POST https://contxt.to/api/contexts \
  -H "Content-Type: application/json" \
  -d '{
    "title": "RFC: Adopting TanStack Query",
    "summary": "Technical proposal to migrate from RTK Query.",
    "content": "## Motivation\n\nOur API layer currently uses RTK Query..."
  }'

Response:

{
  "slug": "rfc-tanstack-kF3mQ",
  "url": "https://contxt.to/s/rfc-tanstack-kF3mQ"
}

See the full API reference for detailed documentation, including AI agent integration patterns and rate limits.

---

Stack

Layer	Choice
Framework	Next.js 16 (App Router)
Database	Turso (libSQL, SQLite-compatible)
ORM	Prisma 7
Auth	NextAuth v5 (Google, GitHub)
Styling	Tailwind CSS v4
UI	shadcn/ui (base-ui)
Fonts	Playfair Display + DM Sans
Testing	Vitest + Playwright
Hosting	Vercel

---

Project Structure

src/
├── app/              # Next.js App Router pages and API routes
│   ├── s/[slug]/     # Shared context pages (browser + AI views)
│   ├── api/          # API routes (contexts, generate-meta, claim, qr, ai)
│   └── page.tsx      # Landing page
├── components/
│   ├── context/      # Context page components (browser + AI)
│   ├── landing/      # Landing page components
│   ├── shared/       # Shared UI (ai-buttons, qr-code)
│   └── ui/           # shadcn/ui components
├── actions/          # Server actions
├── lib/              # Utilities (prisma, auth, schemas, ua, utils)
└── proxy.ts          # AI agent proxy handler (UA-based routing)

---

Contributing

We welcome contributions of all sizes. See CONTRIBUTING.md for:

• Development setup and workflow
• Branch naming conventions
• Code style guidelines
• PR checklist

Check the issues tab for open tickets — good first issues are labelled.

---

Roadmap

• Paste-to-link creation flow
• Dual-render (browser + AI agent) page serving
• AI API proxy with structured YAML output
• Brand and legal pages (Privacy, Terms)
• User dashboard with link management
• Link analytics (views, referrers, devices)
• Link expiration and password protection
• Public API with API keys
• MCP server for direct AI tool integration
• Team / collaboration features

---

License

MIT © RedirHub

---

contxt.to · hello@contxt.to · Discussions