Midjourney,
behind a REST API
Send a prompt over HTTP and get image URLs back. No account to set up, no subscription to hold, no queue to manage. Five cents an image, billed per result.
curl -X POST https://api.journeyapi.dev/v1/images \
-H "Authorization: Bearer $JOURNEY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "a lighthouse on a basalt cliff, fog",
"aspect_ratio": "3:2"
}'What you get
An API key, and nothing else to run.
No account of your own
Nothing to set up, nothing to authorise, nothing to keep alive. An API key is the whole integration.
No subscription
Add funds, spend them when you generate. A quiet month costs nothing. No minimum, no seats, no plan to pick.
99.9% uptime
A pooled, monitored, multi-region backend. Capacity is our problem, not something you have to plan around.
Nothing retained
Prompts and outputs stay yours. Images expire after 30 days, nothing trains a model, nothing is shared.
Pricing
Five cents an image.
You add funds and each result draws from the balance. A job that fails is refunded automatically, so you only ever pay for an image you received.
Batch processing runs the same models at a quarter of the price, without a concurrency guarantee. It suits work that nobody is waiting on.
Get your API key| ImageFrom a prompt, a reference, or both | $0.05 |
| UpscaleFull resolution, 2× or 4× | $0.02 |
| CaptionAn image turned back into prompts | $0.01 |
| VideoOne clip of roughly five seconds | $0.33 |
FAQ
Questions we get asked.
No. You hold no account, no subscription and no seats. One API key is the entire integration, and we issue it within one business day of your request.
Per result, from a prepaid balance. $0.05 an image, $0.02 an upscale, $0.01 a caption, $0.33 a video clip. No monthly fee and no minimum, so a month with no calls costs $0.00.
It is refunded within 60 seconds and you are charged $0.00. That covers screening rejections, render timeouts and our own errors. Cancel a running job and you get the full amount back.
An image returns in 6 to 12 seconds, typically 8. An upscale takes 3 to 5. A five-second video clip takes about 50 seconds. Batch processing costs 25% of the price and lands in 2 to 10 minutes.
60 generation requests a minute, 600 reads a minute, and 8 jobs running at once. Going past 8 queues the job rather than rejecting it. Tell us what you need and we raise the ceiling.
Images stay retrievable for 30 days, then they are deleted. Prompts live only as long as the job record. 0% of your prompts or outputs is used for training, and none of it is shared.
Send the form below and we reply within one business day with a key and your limits configured. Test keys cost $0.00 and return placeholder imagery instantly while you build.
Get in touch
Tell us what you are building.
We onboard teams directly so that rate limits, volume pricing and support are set up before your first call rather than after it.
A real person reads this and replies within one business day.