Pod Engine Python SDK

Official Python client for the Pod Engine podcast intelligence API — search 4M+ podcasts, fetch metadata, transcripts, charts, guest profiles and more.

The client is fully generated from Pod Engine's published OpenAPI specification, so it stays in lockstep with the API and ships typed pydantic models for every request and response.

Install

pip install podengine
# or: uv add podengine

Requires Python 3.10+.

Quickstart (sync)

from podengine import PodEngine

pe = PodEngine(api_key="YOUR_API_KEY")

chart = pe.charts.get_latest_chart(chart_type="apple", country="us", category="top podcasts")
for entry in chart.podcasts:
    print(entry.rank, entry.title)

results = pe.search.search_podcasts(
    search_terms=[
        {
            "searchTerm": "startups",
            "searchType": "text",
            "searchTargets": ["podcast-title"],
            "searchTermOptions": {"matchMode": "optional"},
        }
    ],
)

Get an API key at https://www.podengine.ai/get-started.

Quickstart (async)

import asyncio
from podengine import AsyncPodEngine

async def main() -> None:
    async with AsyncPodEngine(api_key="YOUR_API_KEY") as pe:
        chart = await pe.charts.get_latest_chart(
            chart_type="apple", country="us", category="top podcasts"
        )
        print(chart)

asyncio.run(main())

Configuration

pe = PodEngine(
    api_key="YOUR_API_KEY",
    base_url="https://api.podengine.ai",  # override for staging; or set PODENGINE_API_URL
    source="my-app",                       # sent as the x-source header
    timeout=60.0,                          # HTTPX per-attempt I/O timeout (seconds)
    total_timeout=60.0,                    # overall request budget (seconds)
    max_retries=2,                         # transient 429/5xx/network retries
)

Per-call overrides are available via RequestOptions:

from podengine import RequestOptions

chart = pe.charts.get_latest_chart(
    chart_type="apple",
    country="us",
    category="top podcasts",
    request_options=RequestOptions(timeout=10.0, total_timeout=15.0, max_retries=0),
)

For apps, agents and data pipelines, total_timeout sets a budget for attempts, retry waits and response reads. Both clients default to 60 seconds. Set a positive number on the client or a single call to change it. Set 0 to disable it; client-level None also disables it. In RequestOptions, None inherits the client setting. The existing timeout option keeps its HTTPX per-attempt I/O behavior.

A valid Retry-After sets the minimum retry wait, with 1–251 ms of added jitter. If that wait cannot fit within the remaining budget, the call raises PodEngineConnectionError immediately. It does not retry early. Expired budgets also raise PodEngineConnectionError.

The async client enforces the deadline across the complete call and propagates task cancellation. The sync client caps each HTTPX timeout to the remaining budget and checks the deadline between streamed response reads. It cannot interrupt synchronous I/O already in progress, so a blocking read or custom transport can return after the deadline. No further retry starts after the budget expires.

Error handling

Every failure is a PodEngineError subclass:

from podengine import PodEngine, PodEngineAPIError, PodEngineConnectionError

pe = PodEngine(api_key="YOUR_API_KEY")
try:
    chart = pe.charts.get_latest_chart(chart_type="apple", country="us", category="top podcasts")
except PodEngineAPIError as err:
    print(err.status, err.message, err.request_id)  # non-2xx response
except PodEngineConnectionError as err:
    print("network failure:", err)                  # no HTTP response

License

MIT