switch to gpt-4-turbo by default

release 1.3

Add experimental support for Claude

remove useless code

refactoring for OpenAI and Groq

fix for Groq

the number of pages can now be controlled by the

slight UI redesign

working to fix the config

add a login wall

working on improvements

work on Groq support

fix

fix

anchoring story chunks into the larger length

try to fix the nb of panels issue

improve performance of the LLM

add new "fast LLM" optimization

added suggestion from user + work on the progressive loader

work on a new approach: generate small chunks of a story instead of big one

let's use Zephyr 7b for storytelling

add new turbo mode

fix openai glitch

better captions

minor improvement

fix the dual prompting system

change to the prompt

let's use two pages

seems to be in order

planned maintenance

fixes for the Inference API

increase quality of VideoChain outputs

experimenting with things

the dynamic import is causing issue during build, but it's not an issue we can just put the var init inside the function

fix this error: https://nextjs.org/docs/messages/no-assign-module-variable

fix predict import error and change README

update env variables

fix merge

allow support for Inference API for SDXL

support openai api

Update src/app/queries/predict.ts

new version, able to support Inference API models

fixed upscaling

up

handle a new hallucination case

adding some counter measures to reduce the pressure on the server

various improvements

try to improve robustness of LLM responses

bugfix, improve quality, add layou

add LLM-generated story