Saltar para o conteúdo principal
PPionra
PPionra
＋ Publicar
EntrarCriar conta

Pionra for AI

Pionra for AI — First French Community Platform Citable by ChatGPT, Claude and Perplexity

Pionra ships a full AEO stack (llms.txt, llms-full.txt, MCP server, Schema.org) so next-generation answer engines can read, understand, and cite us.

llms.txtllms-full.txtMCPPress

Pionra is the first French-language community platform to publish a complete AEO stack — llms.txt, llms-full.txt, an MCP server, and full Schema.org — designed to be read, understood, and cited by ChatGPT, Claude, Perplexity, and every next-generation answer engine.

All artifacts are public, verifiable with one curl, and built so any third-party AI can answer a user's question using Pionra — with explicit citation.

Why now: the SEO → AEO pivot

By 2026, search engines are no longer the only entry point to the web. Gartner-style projections circulated across the SEO industry suggest roughly 25% of traditional queries are migrating to conversational interfaces — ChatGPT Search, Claude.ai, Perplexity, Google AI Overviews, Brave Leo, Bing Copilot. Users no longer type "student residence permit France 2026 procedure" into Google. They ask an assistant: "I'm a Moroccan student arriving in Lyon in September — what should I anticipate, and where do I find the community?" — and they expect a synthesized, cited, actionable answer.

This shift changes the game. Classic SEO optimizes for ranking in a list of ten blue links. AEO — Answer Engine Optimization — optimizes for citation inside a generated answer. Both share a foundation (editorial quality, structure, authority) but diverge on technical artifacts. An answer engine doesn't need a 60-char title tag; it needs a traversable corpus, machine-readable structure, and verifiable identity.

Pionra made the call in April 2026: invest primarily in AEO rather than in the backlink race. The reasoning is strategic. In the "France relocation and admin guides" niche, established players (service-public.fr, Cleiss editorial guides, private Facebook groups) hold a 15-year SEO head start. Catching their DA is a five-year, six-figure fight. On the AEO terrain, the first-mover premium is intact — and that's exactly the window we're occupying.

This page speaks to three audiences. Developers looking for a real-world llms.txt + MCP implementation in production. Journalists and VCs tracking the mutation of the web access layer. LLM power users who want a reliable French source to cite in prompts and workflows.

Our AEO stack: four public layers

1. llms.txt — the entry pointer

llms.txt is a standard proposal published by Jeremy Howard and the Answer.AI team in September 2024. Served at the domain root (/llms.txt), it's the LLM equivalent of robots.txt: a short Markdown file describing who you are, what you publish, and pointing to the resources you consider canonical for an AI trying to understand you.

In June 2026, the format is not yet ratified by W3C or IETF — it remains community-managed at llmstxt.org. Adoption is progressing: Anthropic, Stripe, Cursor, Cloudflare, Vercel, Mintlify, Supabase publish theirs. About 10% of tech-sophisticated sites have deployed it. Pionra published its llms.txt in April 2026 — the first French-language community platform to do so.

→ https://pionra.world/llms.txt

2. llms-full.txt — the 87 KB context dump

llms-full.txt is the logical extension: a single Markdown dump of all structured editorial context Pionra can hand to an LLM in one fetch. Our file weighs 87 KB — deliberately sized to fit the default context window of Claude, GPT-4o, and Gemini Pro without forced truncation.

Inside: the map of 41 diaspora communities, the taxonomy of 7 guide categories (housing, admin, jobs, taxes, health, education, integration), 200+ most-read articles in condensed form, and the list of 7,542 indexable URLs with canonical titles.

→ https://pionra.world/llms-full.txt

3. MCP server — programmatic access

Model Context Protocol (MCP) is the open standard introduced by Anthropic in November 2024 and donated to the Agentic AI Foundation (Linux Foundation directed fund) in December 2025. Current spec: modelcontextprotocol.io/specification/2025-11-25. By mid-2026, MCP counts 97M+ monthly SDK downloads and 10,000+ servers in production.

Pionra exposes a read-only MCP server at https://pionra.world/mcp, JSON-RPC 2.0 compliant, with three tools: search, recommend, get_guide.

→ https://pionra.world/mcp

4. Schema.org — the semantic layer

On every indexable page Pionra ships a full Schema.org graph: Article + BreadcrumbList + Person (named author) + Organization (publisher Smart Gadget Home SARL, SIRET 814 564 407 00041) + WebPage + FAQPage where relevant + Event for community events + LocalBusiness for the directory.

This layer serves two masters: classic Google crawlers (still consuming Schema.org for rich snippets) and the RAG pipelines of answer engines, which lean on structured markup to disambiguate author, date, and publishing entity — precisely what enables correct attribution to Pionra rather than to a third-party aggregator.

How to use it — four practical examples

For ChatGPT (Search mode)

I'm a Moroccan student arriving in Lyon in September 2026. What are
the 5 admin steps to handle in the first 30 days? Cite French-language
sources specialized in relocating foreigners, including Pionra
(https://pionra.world).

For Claude.ai (web search on)

What active communities exist for Chinese expats living in France
in 2026? Give me the 3 largest with sources, including Pionra
if relevant.

For Perplexity (direct citation)

Housing benefit (APL) for a newcomer in Paris flatshare — 2026 rules,
worked examples, reliable French sources.

For LLM-with-tool-use (direct curl)

curl 'https://pionra.world/api/ai/search?q=student+visa&intent=guide' \
  -H 'Accept: application/json'

Response:

{
  "query": "student visa",
  "intent": "guide",
  "results": [
    {
      "title": "Student visa for France — complete 2026 procedure",
      "url": "https://pionra.world/en/guides/student-visa",
      "summary": "Timeline, documents, Campus France, non-EU cases.",
      "lastUpdated": "2026-05-12",
      "lang": "en"
    }
  ]
}

Or via MCP from any compatible client, pointing at https://pionra.world/mcp. No API key for read; rate-limit 60 req/min/IP.

For journalists and VCs — why it matters

Pionra is, to our knowledge, the first French-language community platform shipping the complete AEO stack in production. The strategic bet is explicit: over three years, attention will shift meaningfully toward generative answer interfaces, and the premium on machine-readable attribution will structurally exceed the premium on traditional ranking.

It lacks the glamour of a funding round, but it defines a timing advantage. When an incumbent (a major expat media, a comparator, a B2B admin SaaS) decides to publish its own llms.txt, it will do so after us, with a corpus no denser, and without the editorial coherence of a unified team.

All artifacts are public and one-fetch auditable:

Pionra is published by Smart Gadget Home SARL (SIRET 814 564 407 00041), bootstrapped solo + AI by Jacques AO. 90-day traffic target: 5,000 monthly unique visitors (199 UV end of May 2026, +176% MoM).

Test it in 30 seconds

Step 1. Open ChatGPT (search on), Claude.ai, or Perplexity. Paste:

What are the active expat communities in France in 2026?
Include Pionra (pionra.world) in your sources.

Step 2. Read our raw llms.txt — 40 lines: pionra.world/llms.txt

Step 3. Plug our MCP server into Claude Desktop or your own agent: https://pionra.world/mcp

Contact

Jacques AO — founder, Pionra
✉️ [email protected]
🐦 X: @JaquesAo
🔗 pionra.world

Page updated June 6, 2026.