By far the most likely.

NIP-01 should force clients to expose REQs through a regular HTTP endpoint (in addition to websockets) where to fall back on.

I implemented that on relay.zap.store. 20 lines of code.

But there's no interest? Or, what is another hypothesis for these data loading issues? nostr:nprofile1qqsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8gprfmhxue69uhhq7tjv9kkjepwve5kzar2v9nzucm0d5hszxmhwden5te0wfjkccte9emk2um5v4exucn5vvhxxmmd9uq3xamnwvaz7tmhda6zuat50phjummwv5hsx7c9z9 nostr:nprofile1qqsyvrp9u6p0mfur9dfdru3d853tx9mdjuhkphxuxgfwmryja7zsvhqppemhxue69uhkummn9ekx7mp0qythwumn8ghj7anfw3hhytnwdaehgu339e3k7mf0qyghwumn8ghj7mn0wd68ytnhd9hx2tch2deau nostr:nprofile1qqsr9cvzwc652r4m83d86ykplrnm9dg5gwdvzzn8ameanlvut35wy3gpz3mhxue69uhhyetvv9ujuerpd46hxtnfduq3qamnwvaz7tmwdaehgu3wwa5kuegpp4mhxue69uhkummn9ekx7mqcu9929 nostr:nprofile1qqsdv8emcke7k3qqaldwv956tstu40ejg663gdsaayuuujs6pknw7jspp4mhxue69uhkummn9ekx7mqpr3mhxue69uhkummnw3ezucnfw33k76twv4ezuum0vd5kzmqpzfmhxue69uhhqatjwpkx2urpvuhx2uc86mqq0


nostr:nevent1qqsfpq4xlqxgzt305w6lt267zm3vskju490f8sr7733utxcx66t3jfcprpmhxue69uhkv6tvw3jhytnwdaehgu3wwa5kuef0qgsph3c2q9yt8uckmgelu0yf7glruudvfluesqn7cuftjpwdynm2gygrqsqqqqqpzjys8e 

 nostr:nprofile1qqsgz084hwhkwk4mrmegd894ta5cpx7f864h0ruvggm9a3j7zja5asgpzemhxue69uhhyetvv9ujuurjd9kkzmpwdejhgqgewaehxw309aex2mrp0yh8xmn0wf6zuum0vd5kzmp0qyg8wumn8ghj7mn0wd68ytnhd9hx2xqhmyl 

 https://github.com/zapstore/ftsrelay 

 Thanks Fran. Going to check this out today.