{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreia7sk7bjbexeavr66on6yla5to3dcd2mh6mm5dhwop2afv6w2ztkm",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mlvoxw62rei2"
  },
  "path": "/t/codex-rate-limits-discussion-thread/1378553?page=17#post_351",
  "publishedAt": "2026-05-15T15:19:40.000Z",
  "site": "https://community.openai.com",
  "textContent": "There are a lot of things that can be done, and should be looked at when it comes to being efficient with Codex.\n\nBefore I was using it for just a small project, now that project’s a workstation, then I didn’t like the workstation and rebuilt the entire thing.\n\nThe consumption rate of codex credits changed throughout that but it had nothing to do with OpenAi…\n\nMy codebase was growing.\nThe amount of context it took to get in and rebuild based off of stuff I had already proven started taking 8-10 bots running at the same time.\n\nFor a while, it was really bad…\n\nNot every prompt and command is the same in terms of usage.\n\nOne thing I wish I understood about the ChatGPT prompts is they tend to be a little on the nanny size, kind of fat…\n\nStart asking for them to lean. Make a good agents.md…\n\nThere are a million things that could have changed for you on May 10th, and not enough real information in your post for anyone to actually be able to guess at.\n\nLong story short… I was blowing through my 5 hour limit in an hour for several days there - on two accounts…\n\nNow I don’t.. because I’m not having to do things that require that much compute…",
  "title": "Codex Rate Limits Discussion Thread"
}