Unable to use Assistants - (WARNING:tornado.access:404)
I am using caddy, but through https://github.com/lucaslorentz/caddy-docker-proxy. The labels are:
which result in the Caddyfile:
So to my knowledge this should work fine. What am I missing?
windmill.example.org {
handle_path /* {
reverse_proxy 172.19.0.13:8000
}
handle_path /ws/* {
reverse_proxy 172.19.0.12:3001
}
tls {
dns cloudflare <snip>
resolvers 1.1.1.1
}
}lsp-1 | WARNING:tornado.access:404 GET /pyright (172.19.0.5) 1.48ms
lsp-1 | WARNING:tornado.access:404 GET /ruff (172.19.0.5) 0.45ms
lsp-1 | WARNING:tornado.access:404 GET /diagnostic (172.19.0.5) 0.45mswindmill_server:
...
labels:
caddy: "windmill.example.org"
caddy.tls.dns: "cloudflare <snip>"
caddy.tls.resolvers: 1.1.1.1
caddy.handle_path: /*
caddy.handle_path.0_reverse_proxy: "{{upstreams 8000}}"
lsp:
...
labels:
caddy: "windmill.example.org"
caddy.tls.dns: "cloudflare <snip>"
caddy.tls.resolvers: 1.1.1.1
caddy.handle_path: /ws/*
caddy.handle_path.0_reverse_proxy: "{{upstreams 3001}}"There are now many ways to pass specific context to the AIChat for full-code apps. You can use the inspector in the preview, select lines in the editor, use @ mentions (which works with datatables, backend runnables and frontend runnables)
rubenf · 6d ago
**Full-code apps local development in alpha with datatables and any local AI agents ** For the most adventurous of you, in 1.599.0 you can build raw apps fully locally and windmill will take care of providing enough contexts for an agent like Claude to use datatables for storage when initialized with `wmill app new` . Use `wmill app dev` in the app folder to generate a live preview that will also allow you to apply datatables migrations generated by the agents to generate any datatable needed. Highly recommend setting datatable in your workspace settings for the full experience. Require both an updated CLI and windmill. https://www.youtube.com/watch?v=hNQCmpehE2w Merry christmas 🎅
rubenf · 7d ago
little surprise to discover on latest 🎅 It's early, it works especially well with AI and the new data tables. We're eager for your feedback
rubenf · 2w ago