spoo

The official Python SDK for the spoo.me URL shortener API.

from spoo import SpooClient

client = SpooClient()
print(client.shorten("https://example.com").short_url)

• Sync and async clients with the same API
• Fully typed: every response is a Pydantic model, py.typed included
• Sign in with Spoo (PKCE device auth) with automatic token refresh
• Automatic, idempotency-aware retries
• Auto-pagination for list endpoints
• Two runtime dependencies: httpx and pydantic

Migrating from py_spoo_url? See MIGRATION.md.

Install

pip install spoo
# or
uv add spoo

Requires Python 3.10+.

Authentication

client = SpooClient()                    # reads SPOO_API_KEY, else anonymous
client = SpooClient(api_key="spoo_...")  # explicit API key
client = SpooClient(api_key="")          # force anonymous even with env set
client = SpooClient(bearer_token=...)    # a JWT, or a callable returning one

Anonymous clients work under anonymous limits. Create API keys from your spoo.me dashboard. Self-hosting? Point base_url (or SPOO_BASE_URL) at your instance's /api/v1.

Async is the same surface with AsyncSpooClient, await, and async for.

Links

from spoo import SpooClient, UrlFilter, UrlStatus, SortBy

client = SpooClient(api_key="spoo_...")

url = client.urls.create(
    "https://example.com",
    alias="mylink",
    password="Secret@123",
    max_clicks=500,
    expire_after="2026-12-31T00:00:00",
    block_bots=True,
    private_stats=True,
)

# Check availability first if you want a precise reason
check = client.urls.check_alias("mylink")
if not check.available:
    print(check.reason)  # taken | format | length | reserved | emoji_policy

# Fetch one link by id, or by its address
link = client.urls.get(url.id)
link = client.urls.get_by_alias("mylink")                    # your base domain
link = client.urls.get_by_alias("mylink", domain="links.acme.com")

# Iterate everything (auto-pagination), or one filtered page
for item in client.urls.list(sort_by=SortBy.TOTAL_CLICKS):
    print(item.alias, item.total_clicks)
page = client.urls.list_page(filter=UrlFilter(status=UrlStatus.ACTIVE, search="docs"))

# Update, toggle, delete
client.urls.update(url.id, long_url="https://example.com/new", max_clicks=0)
client.urls.set_status(url.id, UrlStatus.INACTIVE)
client.urls.delete(url.id)

Bulk operations

Up to 100 ids per call; results are reported per item instead of throwing:

result = client.urls.bulk_set_status(ids, UrlStatus.INACTIVE)
print(result.summary.succeeded, result.summary.failed)
for row in result.results:
    if not row.ok:
        print(row.id, row.error_code)

client.urls.bulk_delete(ids)
client.urls.bulk_set_expiry(ids, "2027-01-01T00:00:00")   # None clears
client.urls.bulk_set_domain(ids, "links.acme.com")        # None = default

Emoji aliases

url = client.shorten("https://example.com", alias="🚀🔥")       # pick your own
url = client.shorten("https://example.com", alias_type="emoji")  # auto-generate

The SDK validates emoji aliases before sending, against the server's own accepted catalogue (fetched once per client and cached). The catalogue is available directly for building pickers:

emoji_set = client.urls.emoji_set()   # ~1170 entries with names and groups

Claim links

Anonymous creates return a one-time claim_token. After the user signs in, the token proves they created the link and transfers ownership, stats included:

anon_url = SpooClient(api_key="").shorten("https://example.com")

result = client.urls.claim(anon_url.id, anon_url.claim_token)
print(result.status)   # claimed | already_yours | invalid

client.urls.claim_many([(id1, token1), (id2, token2)])   # up to 16

Statistics

Account-wide analytics (requires authentication):

from spoo import GroupBy, Metric, StatsFilter

stats = client.stats.query(
    start_date="2026-07-01",
    end_date="2026-08-19",
    group_by=[GroupBy.TIME, GroupBy.COUNTRY, GroupBy.DEVICE, GroupBy.UTM_SOURCE],
    metrics=[Metric.CLICKS, Metric.UNIQUE_CLICKS],
    timezone="Asia/Kolkata",
    filters=StatsFilter(country=["IN", "US"], utm_campaign=["launch"]),
)
print(stats.summary.total_clicks)
for row in stats.metrics["clicks_by_country"]:   # "{metric}_by_{dimension}"
    print(row)

For a single link you own:

stats = client.stats.for_link(url.id, group_by=[GroupBy.TIME])

Public per-link stats work without authentication. Password-protected links take the password in a POST body, never in the URL:

public = client.stats.public("mylink")
public = client.stats.public("mylink", password="Secret@123")

Exports

Same parameters as query(), returned as bytes in csv, xlsx, json, or xml:

from pathlib import Path
from spoo import ExportFormat

data = client.stats.export(ExportFormat.CSV, start_date="2026-07-01")
Path("report.csv").write_bytes(data)

Path("mylink.xlsx").write_bytes(client.stats.export_link(url.id, ExportFormat.XLSX))

Sign in with Spoo

For connected apps: the PKCE device-auth flow gets you user-scoped tokens without handling passwords. Your app must be registered with spoo.me.

client = SpooClient()
pkce = client.oauth.generate_pkce()
state = client.oauth.generate_state()

# 1. Send the user to the consent page
print(client.oauth.authorization_url("my-app", code_challenge=pkce.challenge, state=state))

# 2. Your redirect URI receives code + state; verify state, then exchange
tokens = client.oauth.exchange_code(code, pkce.verifier)
print(tokens.user.email)

# 3. A provider keeps the session fresh (refresh tokens rotate: persist them)
provider = client.oauth.token_provider(tokens, on_refresh=save_to_disk)
user_client = SpooClient(bearer_token=provider)
print(user_client.me().plan)

When the refresh token itself is rejected, calls raise SessionExpiredError: send the user through the flow again. See examples/sign_in_with_spoo.py for the full loop.

Error handling

Errors map to typed exceptions carrying the backend error code:

Network failures raise APIConnectionError / APITimeoutError; a rejected refresh token raises SessionExpiredError. All of them subclass SpooError.

from spoo import RateLimitError, ValidationError

try:
    client.shorten("https://example.com", alias="taken")
except ValidationError as e:
    print(e.error_code, e.message)
except RateLimitError as e:
    print(f"limited, window resets at {e.reset}")

Retries and configuration

Retries (default 2) honor Retry-After and back off exponentially with jitter. GET/PUT/DELETE retry on 408/429/5xx and network failures; POST/PATCH retry only on 429 and 503, where the server provably did no work.

client = SpooClient(
    api_key="spoo_...",
    base_url="https://your-instance/api/v1",
    timeout=30.0,
    max_retries=3,
    default_headers={"X-Request-ID": "..."},
)

Every request carries an X-Spoo-Client: sdk-py/<version> tag; override it via default_headers if you are building a product on top and want traffic attributed to it.

Note on custom domains: domain= parameters work end to end, but custom domains are currently in a limited beta on spoo.me, so most accounts will see 403 until it opens up.

Examples

Runnable scripts in examples/: quickstart, async usage, analytics, URL management, claim links, emoji aliases, and Sign in with Spoo.

Development

uv sync
uv run pytest
uv run ruff check src/ tests/ examples/
uv run mypy --strict src/spoo/

Versioning

Response models tolerate new fields (extra="allow"), so additive API changes never break an installed version. Breaking changes bump the major version.

License

MIT