API Documentation
All CuteDSL services use a single unified endpoint. Authenticate with your API key, specify the service, and send your request.
Authentication
Connect your Solana wallet on the homepage to get your API key.
Once connected, your real API key will automatically replace YOUR_API_KEY in all code examples on this page.
Rotating Your API Key
If your key is leaked, rotate it to get a fresh one. Authenticate with your current API key plus your wallet address — the old key stops working the instant the new one is issued.
curl -X POST https://cutedsl.cc/api/auth/rotate-api-key \
-H "Content-Type: application/json" \
-d '{
"wallet_address": "YOUR_WALLET_ADDRESS",
"api_key": "YOUR_API_KEY"
}'The response contains your new key — store it securely:
{
"success": true,
"api_key": "cutedsl_xxxxxxxxxxxxxxxxxxxxxxxx", // your new key
"user": { ... }
}Your wallet address is public and safe to share; your API key is secret. You can also rotate with one click on the Account page.
Quick Start
All services share a single endpoint: POST /api/service
curl -X POST https://cutedsl.cc/api/service \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"service": "<service_name>",
...service-specific parameters
}'Response Format
All responses follow this envelope:
{
"result": { ... }, // Service-specific response data
"credits_used": 1000, // $CUTEDSL tokens deducted
"credits_remain": 49000, // Remaining balance
"usd_equivalent": 1.00 // USD value of credits used
}Available Services
Z-Image Turbo
Text-to-image generation powered by CuteZImage with fused Triton kernels. Generates high-quality 1024x1024 images in seconds.
Chronos-2
Time series forecasting with CuteChronos2 acceleration. Achieves 27x speedup over baseline with custom Triton kernels.
Supertonic TTS
High-quality text-to-speech powered by text-generator.io Supertonic and Kokoro voices. Natural-sounding speech synthesis with language, speed, and step controls.
Speech-to-Text
Audio transcription powered by Gemma4. Accurately transcribes speech from audio URLs with support for multiple languages.
Gemma4 Chat
Google Gemma 4 multimodal LLM for chat and vision tasks — an efficient mixture-of-experts (~4B active params). Supports OpenAI-compatible message format with text and image inputs.
Image Caption
Automatic image captioning powered by GitBase. Generates descriptive captions for any image URL.
LTX Video
Text-to-video generation using LTX 2.3. Creates ~6 second 1080p videos at 25fps from text prompts.
Flux Schnell
Fast image generation using Flux Schnell via fal.ai. Produces 1024x1024 images with minimal latency.
LoRA Training
Fine-tune models with LoRA adapters. Train custom adapters for Z-Image or Chronos-2 on your own data.