BYOK is live — 1,000,000 free BYOK requests every month, no top-up requiredLearn more

AI model infrastructure

One API for leading AI models

OpenAI- and Anthropic-compatible. 71 text, image and video models from 13 labs, billed from one USD balance, with leading models priced at or below official list prices.

Prepaid, no subscription. Verify your email for $1 of trial credit (7 days, while this month's budget lasts). Card top-ups from $5.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.boostrail.com/v1",
)
response = client.chat.completions.create(
    model="deepseek-v4.1-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.boostrail.com/v1",
});
const completion = await client.chat.completions.create({
  model: "deepseek-v4.1-flash",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(completion.choices[0].message.content);
curl https://api.boostrail.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "messages": [{"role": "user", "content": "Hello"}]
  }'
Coding agents →
71models59 text · 8 video · 4 image
13labsAnthropic, OpenAI, Google, DeepSeek and 9 more
2API formatsOpenAI-compatible · Anthropic Messages
≤ liston leading modelstoken prices at or below official list prices

Request path

One base URL. 71 models behind it.

Change the base URL and the key; the model id picks the model. When a route for that model fails, the request moves to another route of the same model.

Same codeOpenAI and Anthropic SDKs work unchanged; only the base URL and key differ.
Same modelFailover stays on the model you asked for. It never swaps in a different model.
One billPlatform requests on every route are metered in USD against one prepaid balance.

Models

71 models through one interface

Text, image and video models with transparent usage-based pricing in USD.

View all 71 models →
ModelLabInput / 1MOutput / 1MVerified capabilities ⓘ
claude-opus-5.5AnthropicAnthropic$3.80$19.00Streaming
claude-sonnet-5AnthropicAnthropic$1.90$9.50StreamingToolsJSON schema
gpt-5.5OpenAIOpenAI$4.75$28.50Streaming
gpt-5.4OpenAIOpenAI$2.375$14.25Streaming
gemini-3.1-proGoogleGoogle$2.00$12.00Streaming
deepseek-v4-proDeepSeekDeepSeek$0.627$1.881StreamingTools
qwen3.8-maxAlibabaAlibaba$1.90$5.70
kimi-k3MoonshotMoonshot$2.85$15.00Streaming

Showing 8 of 59 text models · View full catalog →

ModelOutput sizePrice
nano-banana-21K (1024×1024)1K (1024×1024)$0.064 / image
nano-banana-pro1K (1024×1024)1K (1024×1024)$0.127 / image
nano-banana-2-lite1K class (aspect chosen by the model, e.g. 1408×768, JPEG)1K class (aspect chosen by the model, e.g. 1408×768, JPEG)$0.037 / image
muse-image1600×1600 (WebP)1600×1600 (WebP)$0.011 / image

Image generation from $0.011 per image · Image models →

ModelLowest tierFrom
happyhorse-1.0720p720p$0.105 / sec
happyhorse-1.1720p720p$0.105 / sec
grok-imagine-video-1.5480p480p$0.085 / sec
wan-2.6T2V 480pT2V 480p$0.043 / sec
wan-3.0T2V 480pT2V 480p$0.050 / sec
seedance-2.5720p (≈ $0.235/sec)720p (≈ $0.235/sec)$11.263 / 1M video tokens
seedance-2.0480p (≈ $0.075/sec)480p (≈ $0.075/sec)$7.368 / 1M video tokens
seedance-2.0-fast480p480p$0.0473 / sec

Video generation from $0.043 per second · Video models →

Pricing

Published prices, billed per token

Billed per token in USD from one prepaid balance. Leading models are priced at or below their official list prices.

Estimate monthly cost →
claude-sonnet-5USD per 1M tokens · in / out
Official list price$2.00 / $10.00
BoostRail$1.90 / $9.50
  • BalancePrepaid in USD. No subscription.
  • Top-upsCard, USDT or Wise transfer (USD).
  • Cached inputLower cache-read pricing where supported.
  • MediaUsage-based image and video pricing.

Agents and SDKs

Built for agents and developers

Each tool keeps its own config format. Copy the block, paste your key, and pick any model id from the catalog.

Setup guides →
Claude CodeAnthropic Messages over /v1/messages
Codex CLIResponses API over /v1/responses
OpenClawProvider block for openclaw.json
ClineOpenAI Compatible provider in VS Code
OpenCodeProvider entry in opencode.json
ContinueModel entry in ~/.continue/config.yaml
AiderStandard OpenAI-compatible env vars
OpenAI SDKAny OpenAI client: change base URL and key
Use BoostRail as the model API for this project.
- OpenAI-compatible base URL: https://api.boostrail.com/v1
- Anthropic Messages base URL: https://api.boostrail.com
- Auth: read the key from the BOOSTRAIL_API_KEY environment variable; never hard-code it.
- Model ids: https://www.boostrail.com/models/ (for example claude-sonnet-5, gpt-5.4, deepseek-v4.1-flash)
- Docs: https://www.boostrail.com/docs/ · Summary for agents: https://www.boostrail.com/llms.txt

Reliability

Checked routes. Automatic failover.

Routes are checked automatically and re-verified against the official model at least once a week. When a route fails, the request moves to another route of the same model.

01 · Requestclaude-sonnet-5POST /v1/messages
02 · Route ARoute unavailableFAILED
03 · Route BSame modelHEALTHY
04 · ResultRequest completed200 OK
Regular checksEach route is compared with the official model at least once a week.
Bring your own keyRequests on your own provider key can fall back to platform routes of the same model when that key fails.
Before the first tokenFailover happens before the response starts streaming, so a reply never switches routes midway.

Start with one API key.

Prepaid, no subscription. Verify your email for $1 of trial credit (7 days, while this month's budget lasts). Card top-ups from $5.