Skip to main content
  • Free
  • No signup
  • Copyable prompt
  • No AI calls

AI Influencer Prompt Generator

Build image prompts for an AI character, formatted for Midjourney V8.x, FLUX.2, Nano Banana or SDXL and SD 3.5. These are templates: the tool does not call an AI model or generate images.

Model and format

Image model

Instagram format

Character

Visual style

years

Adult characters only.

Prompt for Midjourney V8.x

natural-light photograph of a 24-year-old caucasian woman wearing a fitted t-shirt and shorts, sipping a latte at a small table in a sun-drenched cafe, with a calm, composed expression, rim light with gentle fill, realistic skin texture and true-to-life color, shot on a smartphone with candid framing --ar 3:4 --raw --no text, watermark

Parameters go at the end, each after a space. --raw gives more literal, photographic results. On V8.x, keep the character consistent with the Edit model (up to four reference images); --cref and --oref do not work on V8.

Build the character, not just the prompt

The AI Influencers course covers character creation, reference sets, LoRA training, content production and platform rules. Results depend on your workflow; no income or consistency outcome is promised. Start with the free Starter Kit: working files from the programs that open in your browser.

Get the free Starter KitOr see All Access: $99/mo

How each model reads these prompts

  • Midjourney V8.x: describe subject, medium, setting, lighting and mood, then add parameters at the end. V8.2 has been the default since July 24, 2026, and its Edit model replaces --cref and --oref (Midjourney docs).
  • FLUX.2: subject, action, style, then context, most important first. Black Forest Labs' prompting guide says FLUX.2 does not support negative prompts and suggests naming cameras, lenses or film stocks for photorealism.
  • Nano Banana: Google's image generation guide says to describe the scene rather than list keywords, and to phrase exclusions positively, such as "an empty street" rather than "no cars".
  • SDXL and SD 3.5: both accept a separate negative prompt; Stability's API reference calls it an advanced feature.

From prompt to a consistent character

A prompt is a starting point, not a tested production recipe, and prompts alone do not keep a face consistent. A workflow that matches our AI image generation guide:

  1. Generate a neutral hero portrait: plain background, even light, face fully visible. Save the prompt, model and settings.
  2. Build a reference set from it: front, three-quarter and profile views, a few expressions and outfits. Delete any image where the face drifted.
  3. Reuse the set with reference images (Midjourney's Edit model, Nano Banana or FLUX.2 multi-reference editing), or train a LoRA on an open model when references drift.
  4. Change one thing per generation and review every image at full size before posting.
  5. Label realistic AI characters where platforms require it, such as Instagram's AI-generated profile label.

Questions

Which prompt format does each image model need?

Midjourney V8.x takes a description followed by parameters such as --ar 3:4 --raw. FLUX.2 takes natural language with the subject first and no negative prompt. Nano Banana works best with full sentences describing the scene. SDXL and SD 3.5 take a description plus a separate negative prompt.

Will these prompts give me a consistent character?

Not on their own. Use reference images, such as Midjourney's Edit model, Nano Banana or FLUX.2 multi-reference editing, or train a LoRA on an open model, then check every image against your reference set. No method guarantees the same face every time.

Does FLUX.2 support negative prompts?

No. Black Forest Labs says FLUX.2 does not support negative prompts, so the FLUX.2 version describes what you want instead, such as sharp focus rather than no blur.

Do --cref and --oref work in Midjourney V8?

No. V8.2 has been Midjourney's default since July 24, 2026, and its Edit model replaces Character Reference and Omni Reference. --oref still works on V7 and --cref on V6.

Can I use the images commercially?

It depends on the model and plan, not on this free tool. Midjourney requires Pro or Mega for companies earning over $1M a year, FLUX.2 [klein] 4B is Apache 2.0 while the FLUX [dev] models are non-commercial, and SD 3.5's Community License is free below $1M in annual revenue. Check the current terms before you publish.

Want the whole scenario?

The All Access tools take this job further. Open one to see what it does.

More free tools

All free tools