List available open-source AI models on Replicate via a pay-per-call MPP endpoint.
What it does
This endpoint exposes Replicate's model catalog through the Locus MPP (Micropayment Protocol) gateway. By sending a POST request to `/replicate/list-models`, callers receive a paginated list of the thousands of open-source AI models hosted on Replicate — spanning image generation, language models, speech recognition, video, and more. Pagination is supported via an optional `cursor` parameter returned from previous responses.
The endpoint is part of a broader Replicate MPP service that also includes endpoints for running models (`/replicate/run`), retrieving prediction results (`/replicate/get-prediction`), and fetching model details (`/replicate/get-model`). Each call to list-models costs 4,000 base units of pathUSD (6 decimals), which works out to $0.004 per request, settled on Tempo L2. Payment is handled via the MPP charge intent — the gateway issues a 402 challenge on POST, and the caller settles the micropayment before receiving data.
Note that the probe returned 404 on HEAD and GET because this endpoint only accepts POST requests, which is consistent with the OpenAPI spec. The root domain also returns 404, which is normal for Locus MPP gateways that only serve defined API paths. Documentation is available at the Replicate API reference and the Locus MPP skill file.
Capabilities
Use cases
- —Discovering available open-source AI models before running inference via the companion /replicate/run endpoint
- —Building a model browser or search UI on top of Replicate's catalog
- —Programmatically checking which models are available for automated pipeline selection
- —Agents selecting the right model for a task by enumerating options
Fit
Best for
- —AI agents that need to discover and select models dynamically
- —Developers building model selection interfaces without a Replicate API key
- —Micropayment-native workflows that want pay-per-call access to Replicate's catalog
Not for
- —Running model inference (use the /replicate/run endpoint instead)
- —High-volume bulk catalog scraping where per-call costs add up
- —Users who already have a Replicate API key and can call the native API directly
Quick start
curl -X POST https://replicate.mpp.paywithlocus.com/replicate/list-models \
-H "Content-Type: application/json" \
-d '{}'Example
Request
{
"cursor": null
}Endpoint
Quality
The OpenAPI spec provides clear schema, pricing, and endpoint structure. However, the probe did not capture a live 402 challenge on POST (only HEAD/GET were attempted, both returning 404), no response schema or example response is available, and the crawl pages all returned generic 404 JSON. Pricing and intent are documented in the OpenAPI x-payment-info but not independently verified.
Warnings
- —Probe only tested HEAD and GET; the endpoint requires POST per the OpenAPI spec, so liveness via 402 challenge was not confirmed.
- —No response schema is documented — the shape of the model list is unknown.
- —Currency address 0x20c000000000000000000000b9537d11c60e8b50 is assumed to be pathUSD with 6 decimals based on the Tempo/Locus convention; this is not independently verified.
Citations
- —Endpoint path is /replicate/list-models accepting POST with optional cursor parameter, costing 4000 base units via Tempo charge intenthttps://replicate.mpp.paywithlocus.com
- —Replicate hosts thousands of open-source AI models spanning image generation, language models, speech recognition, video, and morehttps://replicate.com
- —API reference documentation available at Replicate docshttps://replicate.com/docs/reference/http
- —Locus MPP documentation for Replicate available at beta.paywithlocus.comhttps://beta.paywithlocus.com/mpp/replicate.md