Watch an LLM answer token by token: a CLI and web UI that intercepts the OpenAI or Anthropic stream, shows per-token confidence and perplexity from logprobs, and rewrites every other token live. Mock provider needs no API key.
version:'3.8'services:eot:build:.ports:-"8888:8888"environment:-OPENAI_API_KEY=${OPENAI_API_KEY}-ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY}volumes:-./data:/datacommand:["--web","--port","8888","What is consciousness?"]