Python SDK (polylingo)

Official Python client for the PolyLingo REST API. It uses httpx and provides both sync and async clients with the same method names.

• PyPI: polylingo
• Source: UsePolyLingo/polylingo-python

For raw HTTP details, see API reference.

Installation

pip install polylingo

Python: >= 3.9

Sync client

import os
import polylingo

client = polylingo.PolyLingo(
    api_key=os.environ["POLYLINGO_API_KEY"],
    base_url="https://api.usepolylingo.com/v1",  # optional; default shown
    timeout=120.0,  # optional; seconds per request (default 120)
)

result = client.translate(content="# Hello", targets=["es", "fr"], format="markdown")
print(result["translations"]["es"])

client.close()

Context manager:

with polylingo.PolyLingo(api_key="...") as client:
    print(client.languages())

Async client

import polylingo

async with polylingo.AsyncPolyLingo(api_key="...") as client:
    r = await client.translate(content="Hi", targets=["de"])

Use await client.aclose() if not using async with.

Method names match the sync client; all network methods are async def.

Methods (sync and async)

health() / await health()

GET /health

h = client.health()
# async: h = await client.health()

languages() / await languages()

GET /languages

data = client.languages()
langs = data["languages"]

translate(...)

POST /translate

r = client.translate(
    content="# Hello",
    targets=["es", "fr"],
    format="markdown",  # optional
    source="en",        # optional
    model="standard",   # optional: "standard" | "advanced"
)
r["translations"]["es"]
r["usage"]["total_tokens"]

batch(...)

POST /translate/batch

b = client.batch(
    items=[
        {"id": "a", "content": "Hello"},
        {"id": "b", "content": "## Title", "format": "markdown"},
    ],
    targets=["de"],
)
b["results"][0]["translations"]["de"]

usage() / await usage()

GET /usage

u = client.usage()

Jobs (client.jobs)

create / await create

POST /jobs. Returns the 202 body (job_id, status, …).

job = client.jobs.create(content=long_md, targets=["de", "fr"], format="markdown")
# kwargs also accepted: client.jobs.create(**{"content": ..., "targets": [...]})

get(job_id) / await get(job_id)

GET /jobs/:id. When status == "completed", responses include translations and usage at the top level.

translate(...) (convenience)

Polls until completed or failed, or until time runs out.

done = client.jobs.translate(
    content=long_md,
    targets=["de", "fr", "es"],
    format="markdown",
    poll_interval=10.0,   # seconds between polls; default 5.0
    timeout=600.0,        # **total** seconds budget; default 1200 (20 min)
    on_progress=lambda pos: print(f"Queue: {pos}"),
)
done["translations"]["de"]

Async:

done = await client.jobs.translate(
    content=long_md,
    targets=["de"],
    poll_interval=2.0,
    timeout=300.0,
)

API statuses: pending, processing, completed, failed.

Exceptions

import polylingo

try:
    client.translate(content="x", targets=["es"])
except polylingo.AuthError as e:
    print(e.status, e.error)
except polylingo.RateLimitError as e:
    print(e.retry_after)
except polylingo.JobFailedError as e:
    print(e.job_id)
except polylingo.PolyLingoError as e:
    print(e.status, e.error)

Async jobs pattern (summary)

1. Manual: jobs.create → loop jobs.get until terminal state.
2. Helper: jobs.translate with poll_interval, timeout, and optional on_progress.

Prefer jobs for very large content where synchronous translate might hit client or server timeouts.

Types

The package ships py.typed. Response objects are plain dict values aligned with the API; use TypedDict-style annotations in your code if desired.

Changelog

0.1.2

• Maintenance: default API base URL is https://api.usepolylingo.com/v1.

0.1.0

• Initial release: sync PolyLingo, async AsyncPolyLingo, full endpoint coverage including jobs.translate polling helper.