Place to run it · Pictures, Video, Voice and speech, Music and sound

Replicate

A pay-as-you-go service for running thousands of AI models by other makers, for pictures, video, speech, music and text. Mostly used by developers through code, with a web playground for trying models.

Go to Replicate’s website

At a glance

Open or closed: Closed
Only usable through the maker's app or API
The hosting service is proprietary; its model-packaging tool Cog is open source. Openness of each hosted model varies.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Licence: Replicate Terms of Service (The maker's own terms)
source · checked 12 Oct 2026 · checked on the maker's own page
Can I sell what I make?: Depends on the model you choose
s.5.1: Replicate gives you its rights in Output, including commercial use, subject to each model's own licence and third-party terms.
source · checked 12 Oct 2026 · checked on the maker's own page
Cost: No free plan.
Plans below.
How you use it: Maker's own API, In your browser
API with Node, Python and HTTP examples; web Playground linked from the footer. You can also deploy your own models with Cog.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Skill needed: Some setup
Inferred: models can be tried on the web, but regular use is through code and an API key.
source · checked 12 Oct 2026 · inferred or from a secondary source; needs review
Made by: Replicate, LLC
Terms s.12.3(b): notice address in San Francisco, California; country not stated outright. Governing law: California.
source · checked 12 Oct 2026 · checked on the maker's own page
Status: Available
No notice of closure or renaming on homepage or terms.
source · checked 12 Oct 2026 · checked on the maker's own page

What you can make with it

A picture, from a description (main use)
A picture, by changing a photo I have
A video, from a description
A video, by bringing a photo to life
A voice or speech, read my text aloud
Music and sound, a full song with vocals
Music and sound, background music

text to imageimage to imagetext to videoimage to videotext to audiotext to textimage to text

Homepage shows image generation, text- and image-to-video, speech and music models; exact tasks depend on the model you pick.

source · checked 12 Oct 2026 · checked, but the page was unclear or partial

Plans and prices

Plan	Price	Includes
Pay as you go: FLUX1.1 [pro]	$0.04 per image	Commercial use: Depends on the model you choose
Pay as you go: FLUX.1 [dev]	$0.025 per image	Commercial use: Depends on the model you choose
Pay as you go: FLUX.1 [schnell]	$0.003 per image	Listed as $3.00 per thousand output images. Commercial use: Depends on the model you choose
Pay as you go: Wan 2.1 image-to-video 720p (wavespeedai)	$0.25 per seconds of video	480p version $0.09 per second. Commercial use: Depends on the model you choose
Pay as you go: Nvidia T4 GPU	$0.000225 per seconds of computer time	Hardware billed per second ($0.81/hr). L40S $0.000975/s, A100 80GB $0.0014/s, H100 and H200 $0.001525/s. Commercial use: Depends on the model you choose

Some plans are not listed here; see the source for the full list.

No subscription; pay per output or per second of hardware time. Page has 'Try for free' buttons but states no free allowance. Prices shown with '$' only; USD assumed. Enterprise volume discounts on request.

source · checked 12 Oct 2026 · checked, but the page was unclear or partial

Rough cost of one standard image: about $0.025 / £0.02 (estimate, worked out by Plainly, not a price the maker publishes). FLUX.1 [dev] as a typical image model: $0.025 per output image as listed. GBP at 0.75 = 0.01875, rounded. Other models cost more or less.

Rough cost of 5 seconds of 720p video: about $1.25 / £0.94 (estimate, worked out by Plainly, not a price the maker publishes). $0.25 per second of 720p output x 5 seconds = $1.25. GBP at 0.75 = 0.9375.

Rights and safety

Trains on your work?: Doesn't train on your work
s.5.2 licenses your data only to produce output and train your own derivative (fine-tuned) models; s.5.5 allows aggregated, anonymised data to improve the service. Model makers' own terms may differ.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial

Models you can run here

FLUX1.1 [pro], FLUX.1 [dev], FLUX.1 [schnell], FLUX.2, Seedream 4.0, Wan 2.1

Other options for “A picture, from a description”

Adobe Express, Adobe Firefly, Adobe Photoshop, Akool, Artbreeder, Bria, Canva AI (Magic Studio), ChatGPT Images, Civitai, Clipdrop, ComfyUI, Creatify, DeeVid AI, DiffusionBee, DomoAI, Draw Things, Dreamina, Dzine, fal, Filmora, FlexClip, FLUX.1 [dev], FLUX.1 [schnell], FLUX.1 Kontext

About this entry

Facts last checked 12 Oct 2026 against the maker’s own pages (each fact links to its source). Entry status: draft. Plain-English summary written by Plainly.