Code language

REST API

Models API

The public model catalog as JSON: names, categories, prices in credits, limits and configuration. No API key needed, which makes it handy for model pickers in your own product.

List: GET /api/v1/models
One model: GET /api/v1/models/{slug}
Auth: none (key optional)
Cost: free

Try it

A real request against the live catalog, straight from this page.

categorylimitfeatured

GET

curl -s "https://railwail.com/api/v1/models?category=image&limit=3"

Public endpoint: no key, no cost. Press “Send request” to see the live JSON.

List models

GET/api/v1/models

Query parameters

category

string

One of: text, image, video, audio, speech_tts, transcription_stt, embedding, code, multimodal, vla_robotics.

provider

string

The infrastructure provider, one of: openai, anthropic, google, replicate, together, mistral, deepseek, elevenlabs, vastai, huggingface, custom. Lab names such as meta are not valid.

featured

boolean

true for featured models only.

include_unavailable

boolean

true adds the models that cannot run right now, with available: false and the reason.Default false

limit

integer

Page size, at most 500. Without it every listed model comes back in one page, because the OpenAI SDKs read only the first page.Default all (≤ 500)

offset

integer

Skip this many models.Default 0

Unknown category or provider

A value outside the lists above answers 400 invalid_request_error with the code invalid_category or invalid_provider; the message lists the allowed values.

Examples

curl -s "https://railwail.com/api/v1/models?category=image&limit=10"

# pip install requests
import requests

res = requests.get(
    "https://railwail.com/api/v1/models",
    params={"category": "image", "limit": 10},
    timeout=30,
)
for model in res.json()["data"]:
    print(model["id"], model["pricing"]["unit"], model["pricing"]["usd"])

const res = await fetch("https://railwail.com/api/v1/models?category=image&limit=10");
const { data, has_more } = await res.json();
console.log(data.map((m) => m.id), has_more);

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["RAILWAIL_API_KEY"], base_url="https://railwail.com/api/v1")

for model in client.models.list():  # every model the API can run, one page
    print(model.id, model.owned_by)

Response

Shape of the answer; values are placeholders.

JSON

{
  "object": "list",
  "data": [
    {
      "id": "<slug>",
      "object": "model",
      "created": <unix seconds>,
      "owned_by": "<infrastructure provider, e.g. replicate>",
      "name": "<display name>",
      "description": "<short description>",
      "category": "image",
      "provider_model_id": "<id at the provider>",
      "image": null,
      "available": true,
      "unavailable_reason": "<only when available is false>",
      "use_instead": "<slug of a model that can run, when there is one>",
      "lifecycle": { "status": "deprecated", "successor_slug": "<slug>" },
      "pricing": {
        "unit": "per_image",
        "usd": <USD of one default call>,
        "credits": <the same in credits>,
        "basis": "<what usd is the price of>",
        "credit_cost_input_per_1k": <credits, legacy>,
        "credit_cost_output_per_1k": <credits, legacy>,
        "credit_cost_fixed": <credits, legacy>,
        "currency": "credits",
        "credit_usd": <USD per credit>
      },
      "capabilities": { "context_window": <tokens | null>, "max_output_tokens": <tokens | null>, "supported_formats": [] },
      "metadata": { "tags": [], "is_new": false, "is_featured": true, "avg_latency_ms": null, "estimated_duration_seconds": null }
    }
  ],
  "has_more": false,
  "total": <number of models in this list>
}

pricing.unit, usd and credits come from the pricing rule the API bills with (1 credit = USD 0.01). unit is one of per_token, per_image, per_video, per_output, per_second_of_output, per_1000_characters or per_gpu_second; token models have usd: null and carry input_usd_per_1m and output_usd_per_1m instead, GPU-time models the estimated seconds and the amount held up front. The credit_cost_* fields are older columns, kept for existing clients. lifecycle appears only for deprecated, removed or duplicate entries. id is the slug you pass as model everywhere else.

Get one model

GET/api/v1/models/{slug}

Everything from the list plus long_description, configuration.input_schema (the model's inputs), configuration.default_params and examples. A provider model id also resolves, but can match several slugs, so use the slug.

bash

curl -s "https://railwail.com/api/v1/models/flux-1-schnell"

The list holds models that can run

By default the list contains only models the API can run right now: they have a price, and they are neither a duplicate catalog entry nor removed by their provider. ?include_unavailable=true adds the rest with available: false, unavailable_reason and, when there is a runnable alternative, use_instead; calling such a model returns 503 model_unavailable. GET /api/v1/models/{slug} answers 200 with the same fields for every slug. With a key, the request needs the read scope (on by default) and counts toward the key's rate limit.