Generative AI

Replicate HTTP API on looot

Replicate HTTP API gives your agent 3 jobs through looot: generate an image from a prompt, generate a video from a prompt, generate music from a prompt. 4 endpoints run today, from $0.003 per result.

For humansFor agents

Category: Generative AI
Access: Runs on looot's platform key: you do not need your own Replicate HTTP API account
Price: from $0.003 per result
Jobs: 3 jobs, 4 endpoints

Choose Replicate HTTP API when

You need to generate an image from a prompt: Replicate HTTP API is the only provider on looot that runs this job today.
You need to generate a video from a prompt: Replicate HTTP API is the only provider on looot that runs this job today.
You need to generate music from a prompt: Replicate HTTP API is the only provider on looot that runs this job today.

Endpoints and jobs

Each endpoint runs through looot today. Where a job has a benchmark page, its link compares every provider on that job.

Generate an image
image.generate
$0.003 per result
87.9% of 124 runs without an error, median 6.7 s
Google Veo 3.1 video
video.generate
$0.40 per result
47.1% of 17 runs without an error, median 37.3 s
Generate music
music.generate
$0.035 per result
not enough runs to rank
Generate a video
video.generate
$0.50 per result
not enough runs to rank

A minimal request

For Generate an image (replicate-models-predictions), $0.003 per result. Replace each <placeholder> with your value and send a new idempotency key per run.

MCP

Call the looot MCP tool "run" with:
{
  "endpointId": "replicate-models-predictions",
  "input": {
    "model_owner": "<model_owner>",
    "model_name": "<model_name>",
    "input": "<input>"
  },
  "idempotencyKey": "<a new id per run>"
}

REST

curl -X POST https://api.looot.ai/v1/runs \
  -H "Authorization: Bearer $LOOOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"endpointId":"replicate-models-predictions","input":{"model_owner":"<model_owner>","model_name":"<model_name>","input":"<input>"},"idempotencyKey":"<a new id per run>"}'

CLI

looot run replicate-models-predictions --input '{"model_owner":"<model_owner>","model_name":"<model_name>","input":"<input>"}'

Response shape

looot has no stored output format for this endpoint yet, so no response shape is shown here. Run it once and read the answer.

How we measure

Run numbers are looot's own runs of each endpoint over the last 30 days: the share that completed without an error, and the median time to answer. An endpoint with fewer than 10 runs in that window shows "not enough runs to rank" and is never ranked. Prices are what looot charges per call or per result, fee included, read from the live catalog; you prepay credit and each call is charged when it runs.

Read 2026-11-10 23:06 UTC.

Give your agent every data provider, one key.

Start your workspace, top up, and paste one prompt into the agent you already use.

Start your workspace See what agents can do