Technical Specifications of mj-turbo-describe
| Specification | Details |
|---|---|
| Model ID | mj-turbo-describe |
| Provider / Family | Midjourney Describe capability, exposed through CometAPI using the platform model identifier mj-turbo-describe |
| Modality | Image-to-text prompt generation |
| Primary Function | Analyzes an uploaded image and returns descriptive prompt suggestions that can be reused for image generation |
| Input Type | Image upload or image URL, depending on client implementation |
| Output Type | Multiple text prompt suggestions derived from the input image |
| Typical Output Count | Midjourneyโs Describe workflow returns four prompt suggestions per image |
| Interaction Pattern | Asynchronous creative prompt generation from visual input |
| Best For | Reverse-prompting, prompt ideation, style discovery, remix workflows, and converting reference visuals into reusable text prompts |
| Platform Access | Available through CometAPI under the model code mj-turbo-describe |
| Upstream Behavior Notes | Midjourney states Describe is an image-to-text tool intended to inspire prompts rather than precisely reproduce the source image, and repeated runs on the same image can yield different suggestions. |
What is mj-turbo-describe?
mj-turbo-describe is CometAPIโs platform identifier for Midjourneyโs Describe-style image-to-text capability. Instead of generating an image from a text prompt, this model works in the reverse direction: it examines an uploaded image and proposes prompt language that captures visual elements, style cues, and creative framing. Midjourney officially describes Describe as a tool that analyzes an image and offers words and phrases that can be used as prompt inspiration.
In practical terms, this makes mj-turbo-describe useful for users who want to turn existing visuals into reusable prompting language. It can help extract stylistic vocabulary, identify compositional patterns, and accelerate experimentation when building prompts for downstream image generation. Midjourney also notes that the suggestions are inspirational rather than exact replicas of the source image, so the model is best understood as a creative prompt generator, not a strict captioning or forensic vision system.
Because repeated Describe runs may produce different prompt sets for the same image, mj-turbo-describe is also well suited to iterative ideation workflows where variation is valuable.
Main features of mj-turbo-describe
- Image-to-prompt conversion: Converts a source image into text prompt suggestions, making it easier to move from visual inspiration to generation-ready prompt text.
- Multiple prompt candidates: Returns four prompt suggestions per image, giving users several creative directions instead of a single rigid description.
- Prompt ideation support: Helps users discover words and phrases for style, mood, composition, and subject matter that they may not have written manually.
- Non-deterministic creative variation: Re-running Describe on the same image can produce a different set of prompt suggestions, which is useful for brainstorming and exploration.
- Reference-driven workflow: Supports workflows based on uploaded images or image links, aligning well with creators who start from sketches, photos, mood references, or prior generations.
- Fast reuse in generation pipelines: Midjourneyโs Describe results are designed to be immediately reused as prompts, which maps cleanly to API-driven creative automation.
- Style discovery: Particularly useful for identifying visual language embedded in a reference image, including aesthetic cues that can seed future generations.
How to access and integrate mj-turbo-describe
Step 1: Sign Up for API Key
To get started, sign up on CometAPI and generate your API key from the dashboard. After that, store the key securely as an environment variable such as COMETAPI_API_KEY. This key will be used to authenticate every request you send to the mj-turbo-describe API.
Step 2: Send Requests to mj-turbo-describe API
Use CometAPI's Midjourney-compatible endpoint at POST /mj/submit/describe.
curl https://api.cometapi.com/mj/submit/describe \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $COMETAPI_API_KEY" \
-d '{
"prompt": "a futuristic cityscape at sunset --v 6.1",
"botType": "MID_JOURNEY",
"accountFilter": {
"modes": ["TURBO"]
}
}'
Step 3: Retrieve and Verify Results
The API returns a task object with a task ID. Poll GET /mj/task/{task_id}/fetch to check generation status and retrieve the output image URL when the task reaches a terminal state.