I stopped hitting Claude's message limit by building a local AI pipeline that does the heavy lifting
…The obvious answer was a full switch to local AI , but I know better than to not pretend that the abrupt transition would hit like a whiplash. Local models have come a…