Developer API

Integrate a commercial-grade, adaptive mixing engine into your platform. Unlike static AI black-boxes, our deterministic DSP measures the audio (Crest Factor, SNR, PLR) and makes real-time engineering decisions to protect your user's groove and dynamics.

Supported Genres

Pass the preset parameter in your JSON body to change how the Idan mix engine processes the audio.

Make a Request

curl -X POST https://wrattygstudio2.zevcloud.app/api/v1/mix \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "mix",
    "preset": "afrobeats",
    "stems": [
      { "url": "https://your-s3-bucket.com/beat.wav", "role": "beat" },
      { "url": "https://your-s3-bucket.com/vocal.wav", "role": "lead" }
    ]
  }'

Featured Artist Routing: If your track has a featured rapper, set their role to "feature". The engine automatically routes their vocals through the Rap preset (slap delay, heavy saturation) while keeping the main singer on the original genre preset.

Why Adaptive DSP beats Black-Box AI

Neural-network mastering tools blindly force an EQ curve onto your audio. Our engine acts like a human engineer: it measures the source and makes logical decisions.

Vocal Tuning API

Our pitch correction engine uses a hybrid Resplice/PSOLA approach. It automatically detects the song key and preserves formants so the vocal never sounds robotic. You can trigger it via the /api/tune endpoint.

# Internal Adaptive Logic (engine.py)
if q.profile == "poor":
    # Protect bad recordings from robotic artifacts
    s.amount *= 0.70
    s.retune_ms = max(s.retune_ms, 45) 

if singer_err < 10:
    # Singer is already in tune, let them keep their natural vibrato
    s.flex_cents = max(s.flex_cents, 14.0)

if voiced_conf_mean < 0.75:
    # Pitch tracker is unsure (whispers/breath), hold back tuning
    s.conf_gate = (s.conf_gate[0] + 0.06, s.conf_gate[1] + 0.04)

JSON Telemetry & Mix Reports

Unlike black-box AI tools, we don't hide our processing. Every render generates a detailed JSON telemetry report. Once a mix is complete, you can fetch the exact engineering decisions our engine made by calling the report endpoint:

GET /api/v1/report?id=YOUR_JOB_ID

The response includes the full DSP chain, allowing you to display exactly what was done to the user's track (e.g., "The engine cut 2dB at 2500Hz to remove harshness and backed off the loudness to protect your transients").

{
  "plr_guard": {
    "requested_lufs": -9.9,
    "applied_lufs": -10.2,
    "reason": "transient_protection"
  },
  "eq_db": {
    "31": -3.0,
    "80": 1.27,
    "4000": 1.55
  },
  "vocal_chain": {
    "lead": [
      ["highpass", 85.0],
      ["body-mud 320", -3.69],
      ["compress", "opto 2.5:1 @ -20.8"],
      ["saturate", 3.38]
    ]
  },
  "space": {
    "plate": { "rt60": 0.78, "predelay_ms": 38.7 },
    "ducking": { "while_singing_db": 10.6 }
  }
}

Under the Hood: The Adaptive DSP Architecture

Unlike commercial AI mastering tools that use static, black-box neural networks, WrattyGstudio is a Deterministic Adaptive Expert System. It measures the source audio and makes real-time engineering decisions. Here is the complete list of live, active modules running in our engine: