Railwail: The Best AI Models for Image, Video, Voice and Chat

The leading labs

Models from the leading labs

32 developers with at least one model you can run right now, from one account with one API key.

All models

Always the best model

The leading models for every task

When a better model comes out, we add it. These are our current picks per task; the full catalog has many more.

Leaderboard All models

AI-generated
Nano Banana 2
Google DeepMind
Google's fast image model with conversational editing and consistent characters.
Try Nano Banana 2
AI-generated
Flux 1.1 Pro Ultra
Black Forest Labs
FLUX 1.1 Pro in ultra mode: images up to 4 megapixels, with a raw mode for photorealism.
Try Flux 1.1 Pro Ultra
AI-generated
Ideogram V4 Quality
Ideogram
Ideogram v4 at its highest quality: realistic images, creative designs, consistent styles.
Try Ideogram V4 Quality
AI-generated
Recraft V4.1
Recraft
Recraft's latest model, built around design taste: precise prompts and text in the image.
Try Recraft V4.1

Use cases

What will you create?

Real results from the models on railwail, sorted by what people make with them. Hover a tile to see the prompt, click to try the model.

Product photos, packshots and ad visuals for your shop and campaigns.

Text-to-image models Image editing

Watercolour, anime, comics, isometric scenes: any style from one prompt.

Text-to-image models

Nano Banana 2: An isometric 3D miniature diorama of a cozy Japanese ramen shop at night. Warm light spills from the entrance. Tiny detailed figures sit at the counter. Stea…
Recraft V4.1: A brutalist concrete library interior at golden hour — shafts of amber light cut through narrow slit windows and fall across tiered balconies lined with thou…
P Image: Two women in flowing pastel lavender and pale blue dresses, holding hands while walking on a beach at sunset, romantic impressionist oil painting with thick …
Recraft V4: A nice modern vector illustration of a man levitating with a number of various random objects around him. Distinct style, distorted proportions and unique vi…
Fibo: A hyper-detailed, ultra-fluffy owl sitting in the trees at night, looking directly at the camera with wide, adorable, expressive eyes. Its feathers are soft …
Lucid Origin: A fine art painting depicting a serene landscape, bathed in the golden light of sunset. The scene features a rolling hill dotted with wildflowers, a lone tre…
Seedream 4.5: an abstract modern art painting of glass fruit, the glass shines with an iridescent flare, the fruit is transparent
Hunyuan Image 2.1: A cartoon consisting of four panels, arranged in a 2x2 grid, depicts a chameleon's dilemma. Top left: A cartoon-style chameleon rests on a broad, emerald gre…
Stable Diffusion 3.5 Medium: a captivating anime-style illustration of a woman in a white astronaut suit. She has long, dark wavy hair. Surrounding the astronaut are vibrant orange flowe…
Kuaishou Kolors: Expressive illustration of a pug puppy on a Water Slide basked in summer heat
Ideogram V2 Turbo: An illustration of a gold running shoe with the text "Run AI with an API" written on the shoe. The shoe is placed on a pink background. The text is white and…
AuraFlow v0.3: a stylized cartoon image, yellow background, a smiling turtle holding a pizza, the turtle is wearing a white and brown cap, large eyes and a friendly express…

Image editing

Edit any image with words

Upload a picture, say what should change. Drag the handle to compare the original with the result; every pair is a real example of the model named under it.

All image editing models

OriginalEdited

AI-generated

Prompt

remove the car

Edited with FLUX.1 Kontext [dev]Black Forest Labs

Try this edit

AI Flows

Chain models into one flow

Type your idea, pick what you want to get and watch the chain: a prompt writer, an image model, a video model. Open it in the editor, change any step and run it again and again.

All templates and your flows

Your idea

01
Your idea
A cozy reading nook on a rainy evening
02
Claude writes the image prompt
Claude Sonnet 5
Waiting for the step before…
03
Image (FLUX.1 schnell)
FLUX.1 [schnell]

Runs in your account: every step is a normal, billed run and you see the cost before it starts. Change models and steps in the editor.

Build this flow

Video

Made with the latest video models

Veo, Kling, Seedance, Wan and more: from a sentence or a still image to a clip, many with sound.

All video models

Voice cloning

Clone a voice in three steps

Record a few seconds of your own voice or upload a clip, type any text and hear it spoken in that voice. It runs as a normal, billed model run in your account.

All voice cloning models

01Record or upload
10 to 30 seconds of your own voice, right in the browser or as a file.
02Type your text
Anything you want the voice to say.
Hi! This is my cloned voice, made from a short sample.
03Hear your voice
The model speaks your text in the cloned voice.

Only clone your own voice or a voice you have permission to use. You confirm this for every voice sample. Terms of Service

Clone your voice

Try it

Make something right here

One free run on selected fast models (FLUX.1 [schnell], DeepSeek V4.1 Flash), no account needed.

What do you want to make?

Try

Free run, no sign-up.

FLUX.1 [schnell]: a minimalist mountain landscape at golden hour, flat vector illustration — Example · FLUX.1 [schnell]

Open the full FLUX.1 [schnell] playground

Prefer code?API MCP

How it works

Three steps from sign-up to result

Sign up, pick what you need, run it where you work: in the browser, from your code or from an agent.

01
Sign up
With Google, GitHub or email. One account for every model.
Create account
02
Pick a model or a flow
Image, video, voice, chat and more, or chain several models into one AI Flow.
Browse models
03
Run it
In the browser, via the OpenAI-compatible API, or from an agent over MCP.
Docs MCP server

Everything in one place

Every kind of AI model, one account

Images, video, voice, music, text, code, vectors and SVG: one balance, one API key, one dashboard. Open-source models right next to the big labs.

Browse all models

An isometric 3D miniature diorama of a cozy Japanese ramen shop at night. Warm light spills from the entrance. Tiny detailed figures sit at the counter. Steam rises from bowls of r

Text to image

FLUX, Nano Banana, Seedream, Ideogram, Recraft and more.

Text & image to video

Veo, Kling, Seedance, Wan and more, with or without sound.

Just added

67 new models in the last 14 days

Recently added to the catalog and ready to run. The date is when a model joined Railwail, not when its lab released it.

All new models

Open models

Open models, no lock-in

Open-weight models run right next to the closed ones from the big labs. Switch models without rebuilding anything: one OpenAI-compatible API, one key. And because the weights are public, you can also run them yourself.

Browse all models

Also open:Stable Diffusion 3.5 Medium· Stability AI Community License Stable Diffusion XL· CreativeML Open RAIL++-M Qwen-Image-Edit· Apache-2.0 Kokoro TTS 82M· Apache-2.0 Chatterbox· MIT Suno Bark· MIT Stable Audio Open 1.0· Stability AI Community License Whisper· MIT

License names as published by each developer on the model card. Check the license before commercial use.

Compare

Same prompt, different models

Drag the slider to see how two models read the same prompt. Then run it yourself, or put any two models side by side.

Compare any two models All image models

FLUX 1.1 Pro: a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field

FLUX.1 [dev]: a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field

FLUX 1.1 ProFLUX.1 [dev]

AI-generated image

Same prompt

a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field

Compare both models

Leaderboard

Top 3 by quality, price and value

Quality from blind pairwise votes on the public Arena, prices from our pricing rules, value as quality per dollar. Switch the sort, then open the full ranking.

Full leaderboard

Rank 1Best value
Gemini 3.1 Pro
Google DeepMind
Arena score
1,487
±3 · 106,951 votes
Price
$5.40 / 1M tokens
Value
8 pts/$
Rank 2
GPT-5.5
OpenAI
Arena score
1,476
±4 · 66,317 votes
Price
$13.50 / 1M tokens
Value
3 pts/$
Rank 3
GPT-5.4
OpenAI
Arena score
1,466
±4 · 63,526 votes
Price
$6.75 / 1M tokens
Value
3 pts/$

Quality: Text Arena by LMArena, scores published Sep 13, 2026arena.ai leaderboardDatasetCC BY 4.016 of 30 ratedPrices checked Sep 24, 2026

Full ranking: Text & chat

For AI agents

Your agent, every model

Connect Claude Code, Codex or any MCP client once. Your agent then searches the live catalog, compares prices and, with an API key, runs the models itself.

MCP server

Live over MCP · no key

Describe the task. The catalog answers like it answers your agent.

This box calls the real tool recommend_model on our MCP server: a shortlist of models that can run right now, with the tool's reasons and prices.

Try

tools/call recommend_model

waiting for a task

{ "recommendations": [ … ], "no_match": false }

The answer appears here, exactly as your agent receives it: models, prices, reasons.

Same call, same answer as in Claude Code, Codex or Cursor.All tools

Works with Claude Code, Codex CLI, Cursor, Claude Desktop, VS Code and Gemini CLI.
Without a key: search the catalog, compare models and estimate what a workload costs.
With an API key: the agent runs chat, image, video and speech models, billed like the API.

Connect in one line

Pick your client, copy, done

MCP reference

Terminal

claude mcp add --transport http railwail https://railwail.com/api/mcp

1Run the command in your project folder. Claude Code keeps the server for this project.
2Add --scope user to make the server available in every project.
3Check it: ask your agent “Which Railwail image models cost the least per image?” It should call search_models.

Official setup docs

Without a key the server only reads catalog, prices and docs. Nothing to sign up for.

What a session looks like

The user asks, the agent calls a tool, the catalog answers

The tool results below came from the real tools when this page was rendered. Press “Run again live” to repeat the exact call from your browser.

Which image models can draw a logo with clean lettering, and what does one run cost?

tools/call search_models

{
  "query": "logo",
  "category": "image",
  "limit": 3
}

← resultlive catalog data

{
  "matched": 8,
  "catalog_size": 233,
  "excluded_unavailable": 0,
  "models": [
    {
      "slug": "logoai-replicate",
      "name": "LogoAI (SDXL Logo Generator)",
      "provider": "replicate",
      "category": "image",
      "available": true,
      "price": {
        "unit": "per_gpu_second",
        "credits": 3.72,
        "usd": 0.0372,
        "basis": "Estimate for a typical run of about 31.79 GPU-seconds; the real GPU time is billed.",
        "usd_per_unit": 0.00117,
        "estimated_gpu_seconds": 31.79,
        "hardware": "L40S",
        "default_call_usd": 0.0372,
        "default_call": "Estimate for a typical run of about 31.79 GPU-seconds; the real GPU time is billed."
      },
      "context_window": null,
      "avg_latency_ms": null,
      "url": "https://railwail.com/en/models/logoai-replicate"
    },
    {
      "slug": "ideogram-v2",
      "name": "Ideogram v2",
      "provider": "replicate",
      "category": "image",
      "available": true,
      "price": {
        "unit": "per_image",
        "credits": 9.6,
        "usd": 0.096,
        "basis": "One API call: 1 output at the provider defaults",
        "usd_per_unit": 0.096,
        "default_call_usd": 0.096,
        "default_call": "One API call: 1 output at the provider defaults"
      },
      "fixed_price_usd": 0.096,
      "context_window": null,
      "avg_latency_ms": null,
      "url": "https://railwail.com/en/models/ideogram-v2"
    },
    {
      "slug": "recraft-20b-svg-replicate",
      "name": "Recraft 20B SVG",
      "provider": "replicate",
      "category": "image",
      "available": true,
      "price": {
        "unit": "per_image",
        "credits": 5.28,
        "usd": 0.0528,
        "basis": "One API call: 1 output at the provider defaults",
        "usd_per_unit": 0.0528,
        "default_call_usd": 0.0528,
        "default_call": "One API call: 1 output at the provider defaults"
      },
      "fixed_price_usd": 0.0528,
      "context_window": null,
      "avg_latency_ms": null,
      "url": "https://railwail.com/en/models/recraft-20b-svg-replicate"
    }
  ],
  "catalog_url": "https://railwail.com/en/models"
}

8 image models in the catalog match "logo" and can run right now. The first 3:

• LogoAI (SDXL Logo Generator) (logoai-replicate): ≈ $0.0372 per run (billed on GPU time)

• Ideogram v2 (ideogram-v2): $0.096 per image

• Recraft 20B SVG (recraft-20b-svg-replicate): $0.0528 per image

GPU-time prices are estimates; the run is settled on the real GPU seconds.

Same call, anonymous, free.

For developers

One API for every model

OpenAI-compatible: keep the OpenAI SDK and change the base URL.
One API key for chat, image, video, voice and more.
MCP server: agents search the catalog and, with an API key, run models.

Get an API key

Docs MCP server

model: deepseek-v4-1-flash · DeepSeek V4.1 Flash

curl https://railwail.com/api/v1/chat/completions \
  -H "Authorization: Bearer $RAILWAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-1-flash",
    "messages": [
      {
        "role": "user",
        "content": "Explain what a vector database is in two sentences."
      }
    ],
    "max_tokens": 1024
  }'

import os
from openai import OpenAI

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

completion = client.chat.completions.create(
    model="deepseek-v4-1-flash",
    messages=[
        {
            "role": "user",
            "content": "Explain what a vector database is in two sentences.",
        },
    ],
    max_tokens=1024,
)

print(completion.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.RAILWAIL_API_KEY,
  baseURL: "https://railwail.com/api/v1",
});

const completion = await client.chat.completions.create({
  model: "deepseek-v4-1-flash",
  messages: [
    {
      role: "user",
      content: "Explain what a vector database is in two sentences."
    }
  ],
  max_tokens: 1024
});

console.log(completion.choices[0].message.content);

// npm install railwail
import railwail from "railwail";

const rw = railwail(process.env.RAILWAIL_API_KEY);

const res = await rw.chat("deepseek-v4-1-flash", [
  { role: "user", content: "Explain what a vector database is in two sentences." },
], { max_tokens: 1024 });

console.log(res.choices[0].message.content);

Set your key as RAILWAIL_API_KEYCreate API key

model: flux-1-schnell · FLUX.1 [schnell]

curl https://railwail.com/api/v1/images/generations \
  -H "Authorization: Bearer $RAILWAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flux-1-schnell",
    "prompt": "A lighthouse on a rocky coast at sunset, photorealistic"
  }'

import os
from openai import OpenAI

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

image = client.images.generate(
    model="flux-1-schnell",
    prompt="A lighthouse on a rocky coast at sunset, photorealistic",
)

print(image.data[0].url)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.RAILWAIL_API_KEY,
  baseURL: "https://railwail.com/api/v1",
});

const image = await client.images.generate({
  model: "flux-1-schnell",
  prompt: "A lighthouse on a rocky coast at sunset, photorealistic"
});

console.log(image.data[0].url);

// npm install railwail
import railwail from "railwail";

const rw = railwail(process.env.RAILWAIL_API_KEY);

const res = await rw.image("flux-1-schnell", "A lighthouse on a rocky coast at sunset, photorealistic");

console.log(res.data[0].url);

Set your key as RAILWAIL_API_KEYCreate API key

# Catalog, prices and docs, no key needed
claude mcp add --transport http railwail https://railwail.com/api/mcp

# With an API key: agents can also run models
claude mcp add --transport http railwail https://railwail.com/api/mcp \
  --header "Authorization: Bearer $RAILWAIL_API_KEY"

Without a key the server is read-only. With an API key, agents run models over MCP, billed like the API.

Trust

Built in Germany, made for the EU

Railwail is run from Germany under German law. Every point below links to the page that backs it.

Imprint

FAQ

Questions about Railwail

Short answers. Prices and limits are on the pricing page, everything else in the docs.

Which models can I use?

Image, video, voice, music, chat and code models from the leading labs and from open-source teams. When a better model comes out, we add it, so you do not have to switch platforms. Every model has its own page with details and its price.

Can I try it without an account?

Yes. Every browser gets one free run on selected fast models (FLUX.1 [schnell], DeepSeek V4.1 Flash) in the console on this page. Video and voice models need an account.

How does pricing work?

Pay as you go: you top up a prepaid balance and each run is charged from it. There is no subscription, every model shows its price before you run it, and failed runs are refunded. Details are on the pricing page.

Can I chain models together?

Yes, with AI Flows: connect steps such as prompt, image and video, so the output of one model becomes the input of the next. Each step runs like a normal run.

Is there an API?

Yes, an OpenAI-compatible one: use the OpenAI SDK with the base URL https://railwail.com/api/v1 and your Railwail API key. Chat, images, speech, transcription and embeddings work like OpenAI's endpoints; video runs as a job that you poll.

Can AI agents use Railwail?

Yes, through the MCP server at https://railwail.com/api/mcp. Without a key it reads the catalog and the docs; with an API key, agents run models, billed like the API.

Always the best AI model, in one place.

Models from the leading labs

The leading models for every task

Nano Banana 2

Flux 1.1 Pro Ultra

Ideogram V4 Quality

Recraft V4.1

What will you create?

Edit any image with words

Chain models into one flow

Made with the latest video models

Clone a voice in three steps

Make something right here

What do you want to make?

Three steps from sign-up to result

Sign up

Pick a model or a flow

Run it

Every kind of AI model, one account

Text to image

Text & image to video

Image editing

Upscale & remove background

SVG & vector

Voice & text to speech

Music & sound

Speech to text

Chat & reasoning

Code

Embeddings

Robot policies (VLA)

Fine-tuning

67 new models in the last 14 days

GPT-6 Astra

GPT-6 Luna

Nano Banana Pro

Seedream 4

Seedream 4.5

Fibo

Flux Fast

Flux Kontext Max

Open models, no lock-in

FLUX.1 [schnell]

Wan 2.2 Text-to-Video

Qwen Image

Mochi 1

Stable Diffusion 3.5 Large

Wan 2.2 5B Fast

Qwen Image 2512

Stable Diffusion 3.5 Large Turbo

Same prompt, different models

Top 3 by quality, price and value

Your agent, every model

Describe the task. The catalog answers like it answers your agent.

Pick your client, copy, done

The user asks, the agent calls a tool, the catalog answers

One API for every model

Built in Germany, made for the EU

Hosted in Germany

Payment through Stripe

Pay as you go, no subscription

AI marking on generated images

GDPR, controller in Germany

Provider safety filters stay on

Questions about Railwail

Which models can I use?

Can I try it without an account?

How does pricing work?

Can I chain models together?

Is there an API?

Can AI agents use Railwail?

Start with the best models today

What do you want to make?

Text to image

Text & image to video

Image editing

Upscale & remove background

SVG & vector

Voice & text to speech

Music & sound