PoliRural+ — Bridging Rural and Urban Europe · knowledge baseInteractive book · llms.txt · For AI agents

A local GeoAI knowledge base

One district, open geodata and community knowledge, readable by people in a book and by AI agents through open files and an MCP endpoint.

What makes the Klatovsko knowledge base grounded, and how can people and AI agents reach it?

Every factual claim is traceable to a named source or data layer, with date, coverage, method and limits stated. People read and co-write it as an interactive book; AI agents read static pages under /kb/, llms.txt, raw markdown and open data files, or query its MCP endpoint at /api/mcp.

Open in the interactive book · missions, mini-games and memory cards there

The Klatovsko local GeoAI knowledge base: open data layers, grounded blocks, two audiences Open data layers (Copernicus, LPIS, cadastre, OpenStreetMap, national statistics) feed book blocks in which every claim carries a tag pointing to its source layer, and humans review drafts before editorial lift. The same grounded blocks serve two audiences: people through the interactive book, missions and games, and AI agents through static /kb/ pages, llms.txt, raw markdown, open data files and an MCP endpoint at /api/mcp. OPEN DATA LAYERS Klatovsko Klatovy district, Plzeň region CopernicusEarth observation LPISagricultural parcels Cadastreland ownership, boundaries OpenStreetMaproads, rail, services National statisticspopulation, business GROUNDED BOOK BLOCKS Book block id · title · recallQ / recallA Copernicus LPIS Cadastre OSM Statistics each claim → its source layer human review → editorial lift TWO AUDIENCES People Interactive book Missions Games co-author in the author studio AI agents /kb/ static pages llms.txt raw markdown open data files MCP endpoint /api/mcp Same grounded blocks, two front doors; another region rebuilds the stack with its own layers. Book ch. 4, blocks 08–10 · data layers as in D4.1 §2.2, D4.2 §4.1
Open data layers (Copernicus, LPIS, cadastre, OpenStreetMap, national statistics) feed book blocks in which every claim carries a tag pointing to its source layer, and humans review drafts before editorial lift. The same grounded blocks serve two audiences: people through the interactive book, missions and games, and AI agents through static /kb/ pages, llms.txt, raw markdown, open data files and an MCP endpoint at /api/mcp.

A local GeoAI knowledge base: the Klatovsko proof of concept

main tellingexample world: czech-bavariadepth: standard3 min markdown

This chapter’s last part tests a small idea at full scale: a local GeoAI knowledge base for Klatovsko, the Klatovy district in the Plzeň Region of Czechia, which is the Czech side of the Czech–Bavarian pilot. It does not produce another regional profile. It tests whether a book can be the maintained interface between places, evidence and the people deciding about them, and whether AI agents can read the same interface.

The unit: a block tied to a place

Each block answers one local question about an identifiable geography (municipality, catchment, land block, route or habitat) and names the open layers that back it: Copernicus imagery, LPIS, the cadastre, statistics, OpenStreetMap. Blocks state which boundary they use, because datasets often use different ones. The layers themselves are described in what open data says about Klatovsko.

What grounded means

Every factual claim is traceable to a source or layer: publisher, dataset, date, coverage, access link and any transformation. This follows the provenance profile in D4.2, whose priority fields are source, provider, temporal coverage, update status, licence, access level, known limitations and quality status. A mapped overlap is not a cause; satellite greenness is not a diagnosis; local testimony is labelled as testimony. Where evidence is missing, the block says so.

Two front doors

For people, the knowledge base is this interactive book: blocks with a recall question, facets to read by region, depth or language (blocks 08–10 also have Czech tellings), and games. Every block has a ✍️ button that opens the author studio. A draft shared with the group code polirural appears in the consortium gallery (/#trida-polirural), where it collects comments and votes. Editors then lift the best drafts into the served book, and the author’s name travels with the block.

For AI agents, the same content is exposed without a browser:

MCP is the protocol JackDaw and Raven already use (D4.2), and JackDaw lets users connect their own MCP servers at runtime. In principle, the book can therefore become one more tool in a JackDaw conversation.

Why bother

The pilots’ policy brief (D7.3, Annex 10) calls the human effort of curating local data the “hidden infrastructure” of rural AI: months of manual work that is rarely funded. A book makes that work visible, attributed and reusable. D4.2 sets the matching goal for the platform: resources usable by API workflows and AI agents, not only by graphical interfaces.

Success means a reader, human or machine, can move from a practical question to a place-specific explanation, inspect the evidence, suggest a correction, and reuse the pattern in another region.

Místní znalostní báze nad otevřenými geodaty: pokus na Klatovsku

another tellinglanguage: csexample world: czech-bavariadepth: standard3 min markdown

Klatovsko, tedy okres Klatovy v Plzeňském kraji, je česká strana česko-bavorského pilotu projektu PoliRural+. Na něm si ověřujeme jednoduchou věc: dá se z knihy udělat udržovaná místní znalostní báze, ve které se místa, data a lidé, kteří o území rozhodují, potkávají na jednom místě? A dokážou ji číst i AI agenti?

Blok patří k místu

Každý blok odpovídá na jednu místní otázku o konkrétním území (obci, povodí, půdním bloku, trase, biotopu) a jmenuje otevřené vrstvy, o které se opírá: snímky Copernicus, LPIS, katastr, statistiku, OpenStreetMap. Uvádí také, s jakou hranicí pracuje, protože různé datové sady mívají různé hranice.

Co znamená „podložené“

Ke každému tvrzení se dá dohledat zdroj: kdo data vydal, o jakou sadu jde, k jakému datu platí, jaké území pokrývá a jak byla upravena. Vycházíme z profilu původu dat v D4.2. Ten za nejdůležitější považuje zdroj, poskytovatele, časové pokrytí, aktuálnost, licenci, úroveň přístupu, známá omezení a stav kvality. Překryv dvou vrstev ještě není příčina a zelenost ze satelitu není diagnóza. Výpověď místních označíme jako výpověď. Kde důkazy chybějí, blok to řekne.

Dvoje vstupní dveře

Lidé používají tuto interaktivní knihu. Najdou v ní kontrolní otázky, fasety (region, hloubka, jazyk) a hry. Tlačítko ✍️ u každého bloku otevře autorské studio. Koncept sdílený s kódem skupiny polirural se objeví v galerii konsorcia (/#trida-polirural), kde ho kolegové komentují a hodnotí. Nejlepší koncepty redakce převezme do knihy a jméno autora zůstane u bloku.

AI agenti dostanou stejný obsah bez prohlížeče:

MCP je protokol, který už používají JackDaw i Raven (D4.2). JackDaw navíc umožňuje připojit vlastní MCP server za běhu, takže kniha se v principu může stát dalším nástrojem v rozhovoru s JackDaw.

Proč to dělat

Policy brief z pilotů (D7.3, příloha 10) nazývá lidskou práci s přípravou místních dat „skrytou infrastrukturou“ venkovské AI: jsou to měsíce ruční práce, kterou málokdo financuje. Kniha ji zviditelní, podepíše jménem autora a umožní ji použít znovu.

Zdroje:

Related

Sources