Rank 721 in the Majestic Million · checked October 19, 2026
Does pypi.org block AI crawlers?
Partly. As of October 19, 2026, pypi.org's robots.txt blocks none of the 17 AI crawlers we check from the whole site, but keeps all 17 out of parts of it. None of them is named in the file: they all follow its rules for every crawler (User-agent: *).
It has no llms.txt (HTTP 404). Its homepage is 6,982 tokens of raw HTML, 81 as Markdown (1.2%) and 81 pruned for the question “What is pypi.org?”.
Each crawler, by its robots.txt rules
From https://pypi.org/robots.txt (325 bytes, 1 user-agent groups), parsed per RFC 9309. Blocked: no path allowed. Partly blocked: some paths disallowed. The rule counts are those of the groups that apply.
| Crawler | Operator | Verdict | Rules that apply | Disallow rules | Allow rules |
|---|---|---|---|---|---|
| GPTBot | OpenAI | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| ChatGPT-User | OpenAI | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| OAI-SearchBot | OpenAI | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| ClaudeBot | Anthropic | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Claude-User | Anthropic | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Claude-SearchBot | Anthropic | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| anthropic-ai | Anthropic | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| PerplexityBot | Perplexity | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Perplexity-User | Perplexity | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Google-Extended | Partly blocked | Rules for all crawlers (*) | 12 | 0 | |
| Applebot-Extended | Apple | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| CCBot | Common Crawl | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Bytespider | ByteDance | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Amazonbot | Amazon | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| meta-externalagent | Meta | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| cohere-ai | Cohere | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| Diffbot | Diffbot | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
| ContextFetchBot (us) | ContextFetch | Partly blocked | Rules for all crawlers (*) | 12 | 0 |
Its homepage in tokens
https://pypi.org/, fetched once and converted locally with the ContextFetch engine. Tokens are characters divided by four, rounded up, the count the API reports in its X-Raw-Tokens and X-Output-Tokens headers.
| Group | Raw HTML | Markdown | Pruned |
|---|---|---|---|
| Tokens | 6,982 | 81 | 81 |
Read pypi.org through ContextFetch
GET /r/ returns a page as Markdown, with X-Raw-Tokens and X-Output-Tokens headers. Add ?q= and a question to get only the passages that answer it. No key needed for 25 requests a day; a free key has 2,500 credits. pypi.org's robots.txt partly blocks ContextFetchBot: our index and crawls skip what it disallows, and a page you ask for is fetched once on your behalf (see the bot page).
curl -si "https://api.contextfetch.dev/r/https://pypi.org/"curl -X POST https://api.contextfetch.dev/v1/keys/freeHow ContextFetchBot behaves · All sites and the method · Raw data
Give your agent the web
Your free key comes with 2,500 credits, no signup. One credit reads one page.