Block: Context.dev Block

Search the web with natural language, scrape any URL to clean markdown, and crawl entire websites into LLM-ready content through the Context.dev API. Authenticate with an API key and choose one of three operations depending on whether you need web search results, a single-page scrape, or a full-site crawl.

Overview

Property	Value
Type	`context_dev`
Category	`tools`
Color	`#6366F1`

When to Use

Fetch up-to-date information from the web with a plain-English query before passing it to an LLM.
Turn an arbitrary web page into clean, structured markdown for downstream processing or summarisation.
Ingest an entire documentation site or marketing site into LLM-ready content in a single step.
Optionally include or exclude navigation links, headers, and footers from scraped content.
Control crawl scope with page and depth limits to avoid over-fetching large sites.
Monitor API credit consumption automatically via the metadata output.

Configuration

Operation

Selects which Context.dev API endpoint to call. All other fields appear or hide based on this selection.

Label	id
Search	`context_dev_search`
Scrape markdown	`context_dev_scrape_markdown`
Crawl	`context_dev_crawl`

Default: context_dev_search

Query

(Visible when Operation = context_dev_search)

A natural-language question or keyword phrase to send to the Context.dev search endpoint. Accepts {{block.field}} references.

Type: long-input (string)
Placeholder: What is the latest news about...
Required: yes (for Search operation)

Scrape Results to Markdown

(Visible when Operation = context_dev_search)

When enabled, each page returned by the search is additionally scraped and its markdown is included in the result.

Label	id
No	`false`
Yes	`true`

URL

(Visible when Operation = context_dev_scrape_markdown or context_dev_crawl)

The fully-qualified URL to scrape or use as the crawl starting point. Must include http:// or https://.

Type: short-input (string)
Placeholder: https://example.com
Required: yes (for Scrape markdown and Crawl operations)

Main Content Only

(Visible when Operation = context_dev_scrape_markdown)

When enabled, strips headers, footers, and navigation elements from the scraped markdown, returning only the main article body.

Label	id
No	`false`
Yes	`true`

Include Links

(Visible when Operation = context_dev_scrape_markdown)

Controls whether hyperlinks are preserved in the markdown output.

Label	id
Yes	`true`
No	`false`

Max Pages

(Visible when Operation = context_dev_crawl)

Maximum number of pages the crawler will visit, between 1 and 500. Defaults to 100 on the API side when omitted.

Type: short-input (number)
Placeholder: 100

Max Depth

(Visible when Operation = context_dev_crawl)

Maximum link depth from the starting URL the crawler will follow. Set to 0 to fetch only the start page.

Type: short-input (number)
Placeholder: 2

API Key

Your Context.dev API key. Stored as a password field. Reference an environment secret with {{CONTEXT_DEV_API_KEY}}.

Type: short-input (string)
Required: yes (all operations)

Inputs & Outputs

Inputs

id	type	description
`operation`	string	Operation to perform (`context_dev_search`, `context_dev_scrape_markdown`, or `context_dev_crawl`)
`apiKey`	string	Context.dev API key
`query`	string	Search query (Search operation only)
`markdownEnabled`	boolean	Scrape search results to markdown (Search operation only)
`url`	string	URL to scrape or crawl (Scrape markdown and Crawl operations)
`useMainContentOnly`	boolean	Return only main content, excluding headers, footers, and navigation (Scrape markdown only)
`includeLinks`	boolean	Preserve hyperlinks in markdown output (Scrape markdown only)
`maxPages`	number	Maximum pages to crawl (Crawl only)
`maxDepth`	number	Maximum crawl link depth (Crawl only)

Outputs

id	type	description
`data`	json	Result object or array from Context.dev. For Search and Crawl, an array of result objects; for Scrape markdown, an object with `markdown` and `url` keys
`metadata`	json	Response metadata including `creditsConsumed` and `creditsRemaining` for the API key

Tools

Context.dev Search (context_dev_search) — Issues a POST to https://api.context.dev/v1/web/search with a natural-language query and returns an array of search result objects. Optionally scrapes each result page to markdown when markdownEnabled is true. Required params: apiKey, query. Optional: markdownEnabled.

Context.dev Scrape Markdown (context_dev_scrape_markdown) — Issues a GET to https://api.context.dev/v1/web/scrape/markdown for a single URL and returns clean, LLM-ready markdown. Required params: apiKey, url. Optional: useMainContentOnly, includeLinks.

Context.dev Crawl (context_dev_crawl) — Issues a POST to https://api.context.dev/v1/web/crawl and recursively discovers pages starting from a given URL, returning each as markdown. Required params: apiKey, url. Optional: maxPages (1–500, default 100), maxDepth (0 = start page only).

YAML Example

Search

context_dev_1:
  type: context_dev
  name: "Web Search"
  inputs:
    operation: "context_dev_search"
    apiKey: "{{CONTEXT_DEV_API_KEY}}"
    query: "latest AI agent frameworks 2025"
    markdownEnabled: "true"
  connections:
    outgoing:
      - target: agent_1

Scrape Markdown

context_dev_2:
  type: context_dev
  name: "Scrape Page"
  inputs:
    operation: "context_dev_scrape_markdown"
    apiKey: "{{CONTEXT_DEV_API_KEY}}"
    url: "{{trigger.url}}"
    useMainContentOnly: "true"
    includeLinks: "false"
  connections:
    outgoing:
      - target: agent_1

Crawl

context_dev_3:
  type: context_dev
  name: "Crawl Docs Site"
  inputs:
    operation: "context_dev_crawl"
    apiKey: "{{CONTEXT_DEV_API_KEY}}"
    url: "https://docs.example.com"
    maxPages: "50"
    maxDepth: "3"
  connections:
    outgoing:
      - target: knowledge_1

Context.dev Block