Xound Developer API

Programmatic audio & video enhancement — noise removal and studio-grade speech restoration, billed per processed audio-minute.

Processing is asynchronous: start a job, poll its status, then download the result. A job's status moves queued → running → done (or error); progress is 0–100.

Authentication

Every /v1 request needs your secret API key as a bearer token. Create and manage keys in the developer dashboard (passwordless sign-in). Keep the key secret — it grants billable processing.

Authorization: Bearer sk_live_xxxxxxxxxxxxxxxxxxxx

Quickstart

# 1) Start a job — the response includes status_url + result_url
curl https://api.xound.io/v1/enhance \
  -H "Authorization: Bearer $XOUND_API_KEY" \
  -F "file=@input.wav" \
  -F "enhancer=classic" \
  -F "format=wav"

# → {
#     "job_id": "9f2c…", "status": "processing", "is_video": false,
#     "format": "wav",
#     "status_url": "https://api.xound.io/v1/jobs/9f2c…",
#     "result_url": "https://api.xound.io/v1/jobs/9f2c…/result"
#   }

# 2) Poll the status every ~2s until result_available == true
curl https://api.xound.io/v1/jobs/9f2c… \
  -H "Authorization: Bearer $XOUND_API_KEY"
# → { "status": "running", "progress": 80, "result_available": false, … }

# 3) Download the enhanced file
curl https://api.xound.io/v1/jobs/9f2c…/result \
  -H "Authorization: Bearer $XOUND_API_KEY" -o output.wav

Python example

import time, requests

API_KEY = "sk_live_…"
H = {"Authorization": f"Bearer {API_KEY}"}

# Start a job
r = requests.post("https://api.xound.io/v1/enhance", headers=H,
                  files={"file": open("input.wav", "rb")},
                  data={"enhancer": "classic", "format": "wav"}).json()

# Poll until ready
while True:
    s = requests.get(r["status_url"], headers=H).json()
    if s["error"]:
        raise RuntimeError(s["error"])
    if s["result_available"]:
        break
    time.sleep(2)

# Download the result
open("output.wav", "wb").write(requests.get(r["result_url"], headers=H).content)

Enhancers

Pick a model per request with enhancer:

Output formats

Set format for audio inputs (video always returns mp4):

POST /v1/enhance

Start an enhancement job. Returns immediately with a job_id.

Response 200 — includes status_url and result_url so you don't build them yourself.

GET /v1/jobs/{job_id}

Poll this for progress. A job is finished when result_available is true (or error is non-null). 404 if the job isn't owned by your key.

{
  "job_id": "9f2c…",
  "status": "running",          // queued | running | done | error
  "progress": 80,               // 0–100
  "media_duration_s": 42.1,
  "result_available": false,
  "error": null
}

GET /v1/jobs/{job_id}/result

Streams the processed file. Content type follows the request: audio/mpeg, audio/wav, or video/mp4. 404 until ready or after the 3-day retention window.

GET /v1/account

Your account and usage in the current billing period.

{
  "account_id": "…",
  "email": "you@company.com",
  "plan": "starter",
  "usage": { "periodMinutes": 128.4, "totalMinutes": 640.0, "jobCount": 92 }
}

Errors

Billing

Metered by processed audio-minute. Monthly plans include a minute allotment; usage beyond it is billed as overage. See xound.io/api for tiers, or manage your plan in the dashboard.