{
  "id": 7216240,
  "title": "Calling the NovelAI V5 API directly: nai-diffusion-5-full request body, params_version 4/3/1/0 identical, missing v4_prompt is a 500",
  "url": "https://urgent.news/2026/09/14/calling-the-novelai-v5-api-directly-nai-diffusion-5-full-request-body",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-14T01:09:22.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/ilan_kim/calling-the-novelai-v5-api-directly-nai-diffusion-5-full-request-body-paramsversion-4310-133a"
  },
  "original_language": "en",
  "account": "The NovelAI V5 API endpoint behaves inconsistently when directly called, with the request body parameters being flexible but having critical requirements. The params_version field can be set to 4, 3, 1, or 0, and it does not affect the generated image content, which remains pixel-for-pixel identical regardless of the value used. However, omitting either the v4_prompt or v4_negative_prompt field results in a 500 Internal Server Error response instead of the expected 400 Bad Request. Setting the Content-Type manually on a FormData body leads to a 400 error due to a missing multipart boundary. Additionally, the old host URL accepts the generation request but cannot recognize the V5 model ID, resulting in a message indicating an incorrect model ID. The measurements show that the response time is consistent across different params_version values, with an image generation taking around 1.8 to 2.5 seconds. The official web client bundle indicates that the default parameters in the nai-diffusion-5-full branch start with params_version:4, and the code raises stored settings to 4 if params_version is 4. The bundle also suggests that V5 still requires the V4 structure for request parameters, with v5_prompt absent, v4_prompt appearing twice, and v4_negative_prompt being essential. Authentication errors are uniformly represented as 401 Unauthorized, regardless of the header's presence or content.",
  "summary": "nai-diffusion-5-full takes the same body skeleton V4 took, posted to https://image.novelai.net/ai/generate-image , and params_version — the one field every third-party tool fills in differently — returns a pixel-for-pixel identical image whether you send 4, 3, 1, or 0 with this body skeleton. JSON and multipart bodies gave the same result too. What does break it: drop either v4_prompt or…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}