is-crawler 1.3.7

Tiny, zero-dependency crawler detection from user agents.

is-crawler

Tiny, zero-dependency Python library that detects bots and crawlers from user-agent strings. Fast, lightweight, and ready to drop into any web app or API.

Docs & live demo: is-crawler.tn3w.dev

Install

pip install is-crawler

Usage

from is_crawler import (
    is_crawler, crawler_signals, crawler_info,
    crawler_has_tag, crawler_name, crawler_version, crawler_url,
)

ua = "Googlebot/2.1 (+http://www.google.com/bot.html)"

is_crawler(ua)                              # True

crawler_signals(ua)
# ['bot_signal', 'url_in_ua']

info = crawler_info(ua)
# CrawlerInfo(url='http://www.google.com/bot.html',
#             description="Google's main web crawling bot...",
#             tags=('search-engine',))
info.url          # 'http://www.google.com/bot.html'
info.description  # "Google's main web crawling bot for search indexing"
info.tags         # ('search-engine',)

crawler_has_tag(ua, "search-engine")        # True
crawler_has_tag(ua, ["ai-crawler", "seo"])  # False

crawler_name(ua)     # 'Googlebot'
crawler_version(ua)  # '2.1'
crawler_url(ua)      # 'http://www.google.com/bot.html'

API reference

Function	Returns	Description
is_crawler(ua)	bool	Heuristic detection: fast, no DB
crawler_signals(ua)	list[str]	Which heuristic rules matched
crawler_info(ua)	CrawlerInfo | None	url, description, tags: DB lookup for 646 known crawlers
crawler_has_tag(ua, tag)	bool	tag can be str or list[str] (matches any)
crawler_name(ua)	str | None	Product name extracted from the UA string
crawler_version(ua)	str | None	Version extracted from the UA string
crawler_url(ua)	str | None	URL embedded in the UA string

crawler_signals returns a subset of: bot_signal, no_browser_signature, bare_compatible, known_tool, url_in_ua.

crawler_info tags: search-engine, ai-crawler, seo, social-preview, advertising, archiver, feed-reader, monitoring, scanner, academic, http-library, browser-automation.

Middleware example

from is_crawler import is_crawler, crawler_has_tag

@app.before_request
def gate():
    ua = request.headers.get("User-Agent", "")
    if crawler_has_tag(ua, "ai-crawler"):
        abort(403)        # block AI scrapers
    if is_crawler(ua):
        log_crawler(ua)   # track other bots without blocking

How it works

is_crawler: three-step short-circuit, no DB lookup.

1. Positive signal: one fused regex combining bot keywords (bot, crawl, spider, scrape, headless, ...), known tools (playwright, selenium, wget, lighthouse, sqlmap, ...), and URL-in-UA patterns. One hit → crawler.
2. No browser signature: missing WebKit/Gecko/Trident/etc. → crawler.
3. Bare (compatible; ...): classic bot block without OS tokens → crawler.

crawler_signals: exposes which of the five individual checks fired (bot_signal, no_browser_signature, bare_compatible, known_tool, url_in_ua). Useful for diagnostics; is_crawler does not call it.

crawler_info / crawler_has_tag: gated by is_crawler so browser UAs skip the DB entirely. On a crawler hit, patterns (from monperrus/crawler-user-agents plus supplemental tags) are sharded into 48-entry chunks; each chunk's combined filter and its per-pattern regexes compile lazily on first match. Returns url, description, and tags.

Caching: 32k-entry LRU cache on every public function. Repeat UAs hit in ~40 ns.

Regex-free alternative

is_crawler.no_regex, same API, implemented with str.find + char checks. Zero re imports. Useful when embedding in sandboxes that restrict re, when auditing for ReDoS, or when regex engine startup is a concern.

from is_crawler.no_regex import (
    is_crawler, crawler_signals,
    crawler_name, crawler_version, crawler_url,
)

is_crawler("Googlebot/2.1 (+http://www.google.com/bot.html)")  # True
crawler_name("Googlebot/2.1 (+http://www.google.com/bot.html)")  # 'Googlebot'

API is a strict subset of the regex version (no crawler_info/crawler_has_tag those require DB pattern matching). All five functions carry the same 32k-entry LRU cache. Verified against the full fixture corpus (17,043 UAs, 0 mismatches).

no_regex benchmark

Measured against the regex version (Python 3.14, Linux x86_64):

	regex (cold)	no_regex (cold)	speedup
is_crawler	18.7 µs	4.9 µs	3.8×
crawler_url	1.8 µs	0.3 µs	5.8×
crawler_version	2.2 µs	1.5 µs	1.5×
crawler_name	1.7 µs	1.5 µs	1.1×

Cold = cache cleared each iter. crawler_name is slower because the regex version leverages compiled re.sub calls (C-level) to strip comments and browser tokens; char-by-char Python can't match that. The other three win on pure str.find being faster than regex compilation + backtracking. Reproduce with python benchmarks/bench_no_regex.py.

Benchmarks

Measured on Python 3.14, Linux x86_64. Fixture corpus: 1 231 crawler UAs and 15 812 browser UAs. cua is the crawler-user-agents PyPI package (v1.42, no caching).

is_crawler: heuristic detection (no DB)

Corpus	is_crawler	cua.is_crawler	speedup
crawlers only	0.39 µs	61.7 µs	158×
browsers only	3.76 µs	182.7 µs	49×
mixed	0.04 µs	167.7 µs	4000×

Crawler UAs hit a single combined positive regex; browser UAs fall through to a browser-signature check. A 32k-entry LRU cache drives mixed-corpus calls to near-zero amortized cost.

crawler_info / crawler_has_tag: DB pattern lookup

	is_crawler	cua equivalent	speedup
crawler_info	0.53 µs	743.0 µs	1400×
crawler_has_tag	0.14 µs	-	-

Browser UAs short-circuit via is_crawler before touching the DB. Matching walks 48-entry combined chunks to locate the winning pattern in ~1/25 of the full-scan work. crawler_has_tag delegates to cached crawler_info; cost is independent of tag cardinality.

Cold-start

Module	Cold-start	Notes
is_crawler	0.55 ms	JSON parse; regexes stay lazy
crawleruseragents	0.89 ms	JSON parse

DB patterns compile lazily per 48-entry chunk on first match, import and _ensure_db stay cheap.

Single-UA uncached benchmark

	is_crawler	cua.is_crawler	speedup
Googlebot uncached	1.710 µs	70.419 µs	41×

Direct apples-to-apples check on one crawler UA. Same Googlebot/2.1 (+http://www.google.com/bot.html) over 1,000 runs. is_crawler clears caches before each call; cua reloads crawleruseragents before each call.

Formatting

pip install black isort
isort . && black .
npx prtfm

License

Apache-2.0