PromptLens
Get started
Text, image and video models

Prompts written for the model you're actually using.

Claude wants XML. Midjourney wants parameters, not sentences. DeepSeek's reasoning models get worse when you tell them to think step by step. Describe the job once — PromptLens writes it in the dialect of the model you're pasting it into.

Write a promptSee what it costs

10 free every month. No card.

Models it knows

Each one carries its own conventions, its own parameters, and its own list of things that measurably hurt.

Text & reasoning

  • Claude

    Anthropic

    reasoninglong contextcode

  • ChatGPT / GPT

    OpenAI

    reasoningmultimodalcode

  • Gemini

    Google

    long contextmultimodallow cost

  • DeepSeek

    DeepSeek

    reasoningcodelow coststrong in Chinese

  • Grok

    xAI

    real-timereasoning

  • Kimi

    Moonshot AI

    long contextstrong in Chinesecode

  • Qwen

    Alibaba

    strong in Chinesecodelow cost

  • Llama

    Meta

    codelow cost

Image

  • Midjourney

    Midjourney

    stylisedphotoreal

  • ChatGPT Images / DALL·E

    OpenAI

    photorealtext in image

  • Gemini image

    Google

    multimodalfine control

  • FLUX

    Black Forest Labs

    photorealtext in image

  • Stable Diffusion

    Stability AI / open weights

    fine controlstylised

  • Ideogram

    Ideogram

    text in imagestylised

  • Seedream / 即梦

    ByteDance

    strong in Chinesephotoreal

Video

  • Sora

    OpenAI

    cinematicphotoreal

  • Veo

    Google DeepMind

    cinematicmultimodal

  • Runway

    Runway

    fine controlcinematic

  • Kling / 可灵

    Kuaishou

    strong in Chinesefine control

  • Luma Dream Machine

    Luma AI

    cinematicfine control

  • Hailuo / 海螺

    MiniMax

    strong in Chinesephotoreal

How it works

  1. 1

    Pick the model you'll paste into

    Claude, ChatGPT, Gemini, DeepSeek, Grok, Kimi, Qwen — or Midjourney, FLUX, Sora, Veo, Kling. They are not interchangeable, and a prompt tuned for one is a mediocre prompt for another.

  2. 2

    Describe the job in your own words

    A sentence or two is enough. Rough and unstructured is fine, in any of eight languages. The gaps get filled and listed rather than guessed at silently.

  3. 3

    Paste it and get on with the work

    You get the prompt, the system message if that model takes one, the settings worth changing, and the assumptions it had to make — so you can correct the one that's wrong instead of starting again.

Why not just ask a model to write it

A dialect per model, not one template

Structure Claude reads as structure. Parameter syntax Midjourney actually parses. The negative-prompt field where Stable Diffusion does most of its quality work. Written down per model, applied every time.

A checklist per kind of work

A usable video prompt pins down the camera move, the light and what has changed by the end of the clip. A design brief pins down different things. Every generation is held to the list for its domain.

It shows what it guessed

Your brief will be under-specified — they all are. Each gap is filled with a concrete choice and listed as an assumption, so fixing it is one edit rather than a rewrite.

Pricing

Per month, or per generation. Whichever fits how you work.

Free

Enough to find out whether it helps.

$0free

  • 10 generations a month
  • Every model and every domain
  • Saved to your library
  • Briefs up to 2,000 characters
Start free

Pro

Most popular

For anyone writing prompts most days.

$9per month

  • 500 generations a month
  • One brief, up to 3 models at once
  • Briefs up to 12,000 characters
  • Full library and history
  • Markdown export

Studio

For teams shipping prompt-driven work.

$29per month

  • 3,000 generations a month
  • One brief, up to 6 models at once
  • Briefs up to 30,000 characters
  • Priority generation
  • Email support

Not every month

Buy generations outright. They never expire, and they are spent only after the monthly allowance is gone.

50 generations

$5$0.10 each

200 generations

$15$0.075 each

Questions