{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicah6zrv4vawxa2npwiihpzcgx246cma7qzwcg6twbiwmab5uwauy",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mnctg7kyv5w2"
  },
  "path": "/t/june-2026-theme-through-time-chatgpt-api-image-generative-art-gallery-prompt-tips-and-help/1378298?page=67#post_1362",
  "publishedAt": "2026-06-02T13:50:52.000Z",
  "site": "https://community.openai.com",
  "textContent": "It depends on concept.\n\nSort of speak I got a Playground where I have gpt-oss running with ollama, api access, vscode and ChatGPT. For most of the exploration and new creative ideas it’s ChatGPT and vscode. A more mature concepts gets the visuals to a vector store and I have a prompt collector that I use for embeddings. I use API when the workflow is solid and I can apply it in another field.\n\nThe creative ones I like to keep also are documented properly in an Obsidian Vault.\n\nI also have a dedicated workspace for Zero Shot prompting where I use ontology modeling and design to assign prompt resolution to either semantic or visual reasoning.\n\nThis mainly because text models have a tendency to set competing concepts into one clear definition, while the visual reasoning models have a resolution where a fuzzy interpretation can be used.\n\nTo sum up, depends on what I’m doing. Hope it gives a little idea of how I work.",
  "title": "June 2026 (Theme: Through Time) — ChatGPT / API Image Generative Art Gallery, Prompt Tips, and Help"
}