Provider deployment options for Python and TS

[zupdaz]

I’m running into issues trying to deploy create-llama instance to the web. The current structure with the LlamaIndex Server running everything makes deployment quite tricky, and there’s little to no documentation available to guide the process.

I’ve tried test deploying on Vercel and Render. In both cases, the build completes successfully, but I get a “404 Not Found” when trying to access the site. I suspect it’s due to the server setup or missing backend handling.

Could you please provide:
1. Documentation or guidance for deploying Create-Llama to the web
2. Any known working examples or preferred hosting environments

[marcusschiesser]

For NextJS, we're working converting the app to a NextJS project, so coming soon.

For Python, it should be very easy if you use a Dockerfile similar to this:

FROM python:3.11 

WORKDIR /app

ENV PYTHONPATH=/app

# Install Astral uv
# Download the latest installer
ADD https://astral.sh/uv/install.sh /uv-installer.sh

RUN sh /uv-installer.sh && rm /uv-installer.sh

ENV PATH="/root/.local/bin/:$PATH"

COPY . .

# Install dependencies
RUN uv sync

EXPOSE 8000

CMD ["uv", "run", "fastapi", "run"]

Can you test if that works, then we can add it to the template?

[zupdaz]

Great, thanks! I spent a lot of time trying to set up a Next.js project, but the Python Docker version makes things much simpler. That said, the Next.js project, provided by createllama will have additional benefits (like customizability).

[marcusschiesser]

@zupdaz so the Dockerfile did work? Then we can add it to the template

[marcusschiesser]

things to do:

1. add Dockerfile to Python template
2. update README