Curitiba cafes and bars in October 2026: which businesses can AI agents use?
16 cafe and bar and pub websites were scanned (dead addresses left out), and 12 of them could be read. 38% turn an AI crawler away at the edge. Of the ones that could be read, 25% take bookings or orders on their own domain. In all, 1 score A or B. These are the final numbers for the month.
sites scanned
turn an AI crawler away
book or order on their own domain
score A or B
During October 2026, 16 websites of cafes and bars were scanned, with 12 reachable and 4 unreachable from our scanner. Looking at access rules, 6 sites turn at least one AI crawler away at the edge, which is 38% of the scanned sites, while 1 blocks an AI crawler in robots.txt. For menus and pricing, 5 reachable sites present a menu or price list as an HTML page, while 7 provide no menu or price list. Regarding guest services, 3 sites take direct booking or ordering on the business's own domain, 1 routes through an unrecognised third party, and 8 offer no booking or ordering path, with overall grades standing at 0 for A, 1 for B, 3 for C, 4 for D, and 8 for F.
Em outubro de 2026, foram analisados 16 sítios Web de cafés e bares, dos quais 12 estavam acessíveis e 4 inacessíveis. Em relação ao acesso por inteligência artificial, 6 sítios, correspondendo a 38% dos analisados, afastam pelo menos um rastreador no limite da rede e 1 bloqueia um rastreador em robots.txt. No que diz respeito aos conteúdos e reservas, 5 sítios apresentam ementa ou lista de preços como página HTML e 7 não têm ementa ou lista de preços, enquanto 3 sítios, que são 25% dos acessíveis, permitem ao cliente uma reserva ou pedido direto no sítio do business, 8 não têm qualquer via e 1 funciona através de terceiros não reconhecidos. Quanto às avaliações atribuídas aos sítios analisados, registaram-se 0 com A, 1 com B, 3 com C, 4 com D e 8 com F, totalizando 1 sítio com A ou B em conjunto, o que representa 6% dos sítios analisados.
Written by Gemini from the numbers on this page only, 2026-11-05; every number in it is checked against the page before it is shown.
The live index · November 2026 → · what changed · JSON · Markdown
How the sites stood at the end of October 2026
| F | 8 |
| D | 4 |
| C | 3 |
| B | 1 |
| none | 7 |
| HTML page | 5 |
| none | 8 |
| own site | 3 |
| third party | 1 |
| Turn at least one away at the edge | 6 |
| Block one in robots.txt | 1 |
| Unreachable from the scanner | 4 |
Each site is counted once, using its latest full scan before the month ended. The menu and channel counts cover the 12 sites the scanner could read; the AI crawler counts and the grades cover all 16 scanned sites. Dead addresses are left out of every count.
How this page is made
Every business of this kind that has its own website gets the same scan, about once a month, and every scan is kept. This page uses each site's latest scan from before the end of October 2026. A site fixed the month after shows up on the next month's page. A past month doesn't change. The assistant tables use the latest answer per business and question from before the month ended. Methodology · the live index · all cities.