Extraction Adapters
Understand how Spindle chooses the best extraction path for each URL.
Spindle supports multiple extraction strategies so it can handle static sites, article pages, and JavaScript-heavy applications with a single API.
How Extraction Works
When you call /v1/extract, Spindle evaluates the target page and routes the request to the best available adapter.
This gives you cleaner output without needing to decide which scraper or browser workflow to use up front.
Available Adapters
fetch
Best for static and server-rendered pages.
- Fastest path for straightforward HTML pages
- Good for landing pages, documentation sites, and blogs
- Low overhead for high-throughput workloads
readability
Best for long-form written content.
- Optimised for articles and blog posts
- Removes navigation chrome and page clutter
- Returns content that is often ideal for markdown and text outputs
playwright
Best for JavaScript-rendered sites and authenticated flows.
- Handles SPAs and client-rendered content
- Useful when data appears only after hydration
- Best option when simpler adapters cannot fully render the page
Choosing the Right Format
| Format | Recommended use |
|---|---|
markdown | RAG, agents, knowledge bases, documentation pipelines |
text | Lightweight NLP and plain-language processing |
json | Structured downstream processing and custom transformations |
html | Rendering and archiving |
Operational Notes
- Static pages generally return fastest.
- JavaScript-heavy pages may require browser-based rendering and take longer.
- Team quotas and rate limits apply based on your current plan.
Next Steps
- Learn how access is controlled in Multitenancy & RBAC.
- Create secure credentials in API Key Configuration.
- Explore supported request shapes in the API Reference.