Every Tibetan voice deserves a home on the web

The problem

Tibetan civilization is one of the richest and most complex in human history. A living tradition of philosophy, medicine, art, music, and language that has endured for over a thousand years — practiced today by millions of people across India, Nepal, the West, and increasingly in exile communities everywhere.

And yet on the internet — the infrastructure that shapes how the world learns, remembers, and understands — it barely exists.

The reasons are many — occupation, suppression, diaspora, the technical barrier of getting online — but the result is the same. Monasteries with no websites. Teachers whose schedules live in WhatsApp groups. Artists whose work exists physically but nowhere on the web. Writers publishing to algorithms that bury their words.

A 2026 essay by Thupten Chakrishar documented the scale of this crisis with precision — examining not just the absence of Tibetan content, but how Chinese AI systems are actively filling that vacuum with Beijing's narratives, and how the window to shape what AI knows about Tibet is narrowing fast. We encourage you to read it.

The opportunity we can actually take

Training an AI model costs millions of dollars. Lobbying technology companies for Tibetan language support takes years. These are real fights worth fighting, and people are fighting them.

But there is something else — something every monastery, every teacher, every artist, every blogger can do right now. They can put their work on the web.

Common Crawl — the web crawl that feeds virtually every major AI system, including GPT, Claude, Gemini, and Llama — indexes billions of pages from across the internet every few months. When a Tibetan monastery publishes their history, when a thangka painter documents their technique, when a Tibetan writer shares their stories — that content enters the training data that shapes what AI knows about the world for decades.

Every Tibetan website is not just a website. It is a contribution to the permanent record of human culture. It is training data. It is the source material that future AI — and future generations — will learn from.

The barrier was never motivation. It was access. Building a website required technical skills most community members don't have, or thousands of dollars that monasteries and cultural organizations couldn't spare. Jinpa removes that barrier entirely.

How Jinpa helps — including the parts you can't see

In solving this problem, we built something the world had never seen before: the first CMS that runs entirely on GitHub Pages — no servers, no databases, no dependencies. With Jinpa, anyone can go from zero to a professional, live website in 90 seconds. No coding. No cost. No recurring subscription. That's the visible part.

The invisible part is what makes Jinpa different from any other free website tool: every Jinpa site is engineered from the ground up to be discovered, indexed, and understood — by search engines, by AI crawlers, and by the systems that will shape the web for generations to come.

• llms.txt — a new standard designed specifically for AI crawlers. Every Jinpa site automatically generates a clean, structured overview of its content that LLMs can read and index without rendering JavaScript.
• Structured data (JSON-LD) — every page includes machine-readable Schema.org markup. Search engines and AI systems know exactly what the content is, who created it, when it was published, and what language it's in.
• Sitemap — automatically generated and submitted to Google, Bing, and other search engines so every new page is indexed as fast as possible.
• Explicit AI crawler welcome — Jinpa's robots.txt explicitly welcomes GPTBot, ClaudeBot, PerplexityBot, CCBot (Common Crawl), and every other major AI crawler by name. Most sites either block them or ignore them. Jinpa sites invite them.
• Open Graph and RSS — content spreads correctly across social platforms, news aggregators, and research tools, and is attributed properly.
• Static HTML — no JavaScript rendering required. Content is fully readable by every crawler the moment the page is live. A monastery in Mundgod with slow internet gets the same crawl quality as a multi-million dollar company with a full-time engineering team.

A center using an old WordPress install with a slow plugin stack might never get indexed well. A Jinpa site gets everything right by default, automatically, without the site owner ever thinking about it.

What we're building toward

Jinpa 1.0 solves the access problem. What comes next is a platform built specifically for how Tibetan communities create and share content.

• Tibetan font support — Noto Serif Tibetan and Jomolhari available as default options so bilingual content looks beautiful without any configuration.
• Bilingual templates — designed from the start for content in both Tibetan script and English, with proper text direction and layout handling.
• Dharma-specific content types — structured fields for teachings, lineages, practice schedules, teacher biographies, and event calendars.
• More templates — portfolios for artists, cultural archive templates for NGOs, simple family sites, documentation templates for preservation projects.

Who we are

Jinpa is an open-source project built for the Tibetan community, by @vajradog and friends. It runs on GitHub Pages — the free hosting infrastructure that powers millions of sites worldwide — and it will always be free for communities that need it.

The code is open. The platform is open. If you're a developer and want to contribute, everything is at github.com/getjinpa.

The mission is simple: every Tibetan voice deserves a home on the web. The window to build that presence is measured in years, not decades. We're making it free.