IntellectKit provides clean, structured data from the web — article extraction, product data, page metadata, link analysis, and developer utilities. One API key, instant results.
View Endpoints See PricingA lightweight, fast API for extracting structured data from the web — designed to slot directly into AI agent pipelines.
Pull clean article text, author, publish date, images, and a reading-time estimate from any URL.
Extract name, price, currency, brand, availability, and rating from product pages using structured data and heuristics.
Retrieve Open Graph tags, Twitter cards, JSON-LD, canonical URL, favicon, and language in one call.
Get all links from a page, classified as internal or external, with anchor text and rel attributes.
Strip nav, footer, ads, and noise — return just the readable prose with word and character counts.
Email validation and DNS record lookup. More utilities added regularly.
All endpoints under /v1/ require an X-API-Key header.
{"status":"ok"}.Start free. Scale when you need to.
IntellectKit endpoints are designed to be called directly by LLM tool-use — clean JSON, predictable schemas, no scraping headaches.
Every response follows a consistent schema. Null fields are explicit, never missing. Easy to parse with structured outputs.
All data endpoints use GET requests with URL query params — trivial to add as tool definitions in any agent framework.
No database lookups on the critical path for extraction. Fetch, parse, return.
Errors return descriptive JSON with a reason field — agents can relay them directly to users or retry logic.
# Add to ~/.claude/claude_desktop_config.json { "mcpServers": { "intellectkit": { "command": "npx", "args": ["@intellectkit/mcp-server"], "env": { "INTELLECTKIT_API_KEY": "ik_your_key_here" } } } }
import anthropic client = anthropic.Anthropic() tools = [{ "name": "extract_article", "description": "Extract article content from a URL", "input_schema": { "type": "object", "properties": {"url": {"type": "string"}}, "required": ["url"] } }] # When Claude calls the tool, forward to IntellectKit: resp = requests.get( "https://api.intellectkit.dev/v1/extract/article", params={"url": tool_input["url"]}, headers={"X-API-Key": "ik_your_key_here"} )
tools = [{
"type": "function",
"function": {
"name": "extract_text",
"description": "Get clean readable text from any URL",
"parameters": {
"type": "object",
"properties": {"url": {"type": "string"}},
"required": ["url"]
}
}
}]
# Full schemas for all 7 tools in AGENTS.md
One API key replaces a stack of scraping tools, browser dependencies, and per-service integrations.
Article extraction, product data, metadata, links, text, email validation, DNS — all from a single key with a unified auth pattern.
Every field is typed and null-safe. No missing keys, no inconsistent shapes. Feed responses directly into structured output parsers.
Fields like summary, wordCount, and readingTimeMinutes are designed for agent reasoning, not just display.
No Playwright, Puppeteer, or headless Chrome to maintain. Fast, lightweight extraction with zero infrastructure overhead.
Pass your API key in the X-API-Key header on every request to /v1/ endpoints.
Include the header with every API request:
curl https://api.intellectkit.dev/v1/extract/article?url=https://example.com \
-H "X-API-Key: ik_your_key_here"
Python example:
import requests
resp = requests.get(
"https://api.intellectkit.dev/v1/extract/article",
params={"url": "https://example.com/post"},
headers={"X-API-Key": "ik_your_key_here"}
)
data = resp.json()
Missing or invalid keys return 401 Unauthorized.
Requests over your daily limit return 429 Too Many Requests.