tezzinc — Python SDK

The official Python client for the Tezzinc video transcoding & subtitling platform. A thin, typed wrapper over the same REST API the console and the MCP server use.

pip install tezzinc

Quickstart

import os
from tezzinc import Tezzinc, TezzincError

client = Tezzinc(api_key=os.environ["TEZZINC_API_KEY"])   # or just Tezzinc()

# 1. Register a source
asset = client.assets.create(
    source_uri="https://example.com/clip.mp4",
    content_type="video/mp4",
)

# 2. Submit a transcode
job = client.jobs.create(
    asset_id=asset.id,
    spec={"preset_id": "h264_1080p"},
    correlation_id="my-first-job",
)
print("submitted", job.id, "status=", job.status)

# 3. Wait for it (polling; use a webhook on a server)
final = client.jobs.wait_for(job.id, timeout=600, poll_interval=2.0)

# 4. Read the outputs (signed, short-lived URLs)
for out in final.outputs:
    signed = client.outputs.signed_url(out.id, ttl_seconds=900)
    print(out.preset_id, out.byte_size, signed.url)

Configuration

The client reads the API key from, in order: the api_key= argument, the TEZZINC_API_KEY environment variable, or Tezzinc.from_config("config.json"). The base URL defaults to https://tezzinc.com and can be overridden with base_url= or TEZZINC_API_BASE.

Errors

Every non-2xx response raises TezzincError with .status (HTTP code), .code (machine-readable, e.g. "quota_exceeded"), .message, and .headers (read Retry-After on a 429). Transport failures raise TezzincConnectionError (a subclass, status == 0).

from tezzinc import TezzincError

try:
    client.jobs.create(asset_id=asset.id, spec={"preset_id": "h264_1080p"})
except TezzincError as e:
    if e.status == 429:
        retry_after = int(e.headers.get("Retry-After", "60"))
    elif e.status >= 500:
        ...  # retry with jitter

Resources

Responses are attribute-accessible (job.status, job.outputs[0].preset_id) and also expose .raw for the underlying dict.

MIT licensed.