Test PubContext API endpoints directly in your browser. Select an endpoint, configure parameters, and send requests.
Click any example to load it in the playground.
// Response will appear here after sending a request
When you make a request, the API returns several headers that tell you how the response was generated.
Documentation was found in the memory cache (super fast response)
Not in memory cache, had to check disk storage or scrape from pub.dev
Served from memory cache (fastest, ~instant)
Loaded from disk storage at api/data/{package}.json (fast)
api/data/{package}.json
Just scraped live from pub.dev (slowest, typically a few seconds)
Fetched from pub.dev API (used for search results)
From curated package database (categories, related, alternatives)
The API follows this waterfall when fetching documentation:
Cache: MISS + Source: scrapeTakes a few seconds to scrape
Cache: MISS
Source: scrape
Cache: HIT + Source: memoryInstant response from memory
Cache: HIT
Source: memory
Cache: MISS + Source: diskFast, loaded from disk then re-cached
Source: disk
Token count for complete documentation (regardless of depth filter). For detailed estimates, use the /token-estimate endpoint.
Uses format-specific ratios: 4.0 chars/token for Markdown, 4.7 chars/token for JSON. This header only appears for documentation responses.