cyphr-sdk 0.1.1

Official Python SDK for the Cyphr AI mixing API

Cyphr Python SDK

Official client for the Cyphr AI mixing API. Upload stems, run an automated AI mix, and download the finished WAV — all from Python.

Install

pip install cyphr-sdk

Requires Python 3.9+.

Authentication

1. Sign in at cyphrmusic.com.
2. Open Account → API keys and create a key (cyphr_sk_…).
3. Pass it to the client or set CYPHR_API_KEY.

The account page also shows the API base URL to use with base_url= or CYPHR_API_BASE_URL.

Quick start

from pathlib import Path
from cyphr import CyphrClient, StemSpec

client = CyphrClient(api_key="cyphr_sk_…")

mix_path = client.mix_project(
    title="My Song",
    stems=[
        StemSpec("vocals.wav", "vox_lead", Path("vocals.wav")),
        StemSpec("beat.wav", "multiple_instruments", Path("beat.wav")),
    ],
    output_path="mix.wav",
)
print("Downloaded:", mix_path)

Step-by-step workflow

Use this when you want control over uploads or status polling:

from cyphr import CyphrClient

client = CyphrClient(api_key="cyphr_sk_…")

# 1. Create project — returns presigned upload URLs
created = client.create_project(
    title="My Song",
    stems=[
        {"filename": "vocals.wav", "trackType": "vox_lead"},
        {"filename": "beat.wav", "trackType": "multiple_instruments"},
    ],
)

# 2. Upload each stem (PUT to presigned URL; headers must match)
for upload, path in zip(created.uploads, ["vocals.wav", "beat.wav"]):
    client.upload_stem(upload, path)

# 3. Queue for mixing
client.queue_project(created.project.id)

# 4. Wait until status is "completed"
project = client.wait_for_completion(created.project.id)
print(project.status)

# 5. Download the mix WAV
client.download_mix(created.project.id, "mix.wav")

Configuration

Setting	Environment variable	Constructor kwarg
API key	CYPHR_API_KEY	api_key=
API base URL	CYPHR_API_BASE_URL	base_url=

# Explicit configuration
client = CyphrClient(
    api_key="cyphr_sk_…",
    base_url="https://your-api-url.lambda-url.us-east-1.on.aws",
    timeout=60.0,  # HTTP timeout in seconds
)

If base_url is omitted, the SDK uses its built-in production URL when set, otherwise the dev API URL shipped with the package.

Track types

Every stem needs a trackType when creating a project:

Value	Use for
vox_lead	Lead vocal
vox_back	Backing vocals / harmonies
multiple_instruments	Full beat or multi-instrument stem
bass	Bass line
808	808 / sub bass
drum_kit	Full drum mix
kick	Kick drum
snare	Snare / clap
hi_hat_cymbals	Hats / cymbals
keys	Piano / synth keys
guitar	Guitar
sfx	Risers, sweeps, FX
multi_normalize	Already-processed stem (normalize only)

Supported audio formats: .wav, .flac, .mp3, .m4a, .aiff, .aif

API reference

CyphrClient

Method	Description
verify()	Check API key; returns { "userId": "…" }
list_projects()	List AI mix projects
create_project(title, stems, genre=None)	Create project + presigned upload URLs
upload_stem(upload, source)	Upload one stem file
queue_project(project_id)	Start mixing after uploads
get_project(project_id)	Get project status
wait_for_completion(project_id, …)	Poll until done
get_mix_download(project_id)	Presigned download URL
download_mix(project_id, dest=None)	Download WAV to disk
mix_project(title, stems, …)	End-to-end helper

Types

• StemSpec(filename, track_type, source) — local file to upload in mix_project
• Project — id, title, status, mix_status, …
• StemUpload — presigned upload target from create_project
• MixDownload — presigned download URL from get_mix_download

Exceptions

• CyphrConfigError — missing API key or invalid client setup
• CyphrApiError — API returned an error (status_code, body)
• TimeoutError — wait_for_completion exceeded its timeout

In Python: help(CyphrClient) for full method documentation.

Project statuses

Typical flow: uploading → queued → pending → completed

If classification needs manual review in the app, status may become awaiting-review. Failed mixes return failed.

License

MIT