One line of Python lets your agent read any page as structured data — and fires a webhook the moment something meaningful changes. No browser farms. No polling.
No credit card required · 100 free API calls / month
Not just “the page changed.”
“Price dropped 23% (¥129 → ¥99). Stock unchanged.”
Field-level structured diffs — noise-filtered, classified, and delivered as signed webhook events.
Change Events
Point a monitor at a URL. On every check, we diff the new snapshot against the last one, filter out noise, and POST only real changes to your webhook — typed, classified, and ready to act on.
Per-field before/after, not a noisy text dump. Price, stock, headline, status — each tracked independently.
Timestamps, CSRF tokens, analytics IDs, and dynamic attributes are stripped before diffing — 28 patterns out of the box.
Each change is tagged (price_drop, stock_out, copy_change…) with a 0–1 confidence score.
Every payload is HMAC-SHA256 signed. Failed deliveries retry with backoff and land in a dead-letter queue — all visible in delivery logs.
{
"event": "page.changed",
"task_id": "task_abc123",
"url": "https://example.com/product",
"change_type": "price_drop",
"confidence": 0.97,
"diff": {
"price": { "before": 129.00, "after": 99.00 },
"stock": { "before": "in_stock", "after": "in_stock" }
},
"summary": "Price dropped 23% (¥129 → ¥99). Stock unchanged."
}Extract structured data from any URL, or describe a monitor in plain English. 10 free anonymous requests per day.
Extract structured data from a URL
JSON schema for structured extraction. Leave empty for auto-detection.
Read Primitives
The same client that watches for changes also fetches structured data on demand — for one-off agent calls and bulk ingestion.
/v1/extractPass any URL + schema, get typed JSON back. AI auto-detects selectors if schema is omitted.
/v1/searchWeb search with LLM-ready markdown output and extractable links.
/v1/crawlBulk-extract a site via sitemap or link-following. Aggregated structured data.
Python SDK
Install with pip. Call in one line. Extract on demand, or start a monitor that calls back the moment something changes. No browser farms, no proxy management, no polling loops.
import flypython
# 1. Let your agent read any page as structured data
data = flypython.extract(
url="https://example.com/product",
schema={"price": "number", "stock": "string"},
)
# 2. Watch it — get a webhook event when it changes. No polling.
flypython.monitor(
url="https://example.com/product",
schema={"price": "number", "stock": "string"},
webhook="https://myagent.com/events",
schedule="0 */6 * * *", # every 6 hours
)Use Cases
Give your LLM agents reliable web access — and let them react the moment a page changes, via webhook. No browser farms, no polling loops.
Turn web pages into event sources. Build price watchers, job trackers, and research pipelines that trigger the instant something meaningful changes.
Embed live web data into your product. Structured JSON feeds your RAG pipeline; change events drive alerts, digests, and agent actions.
Pricing
Usage-based, no subscriptions. 100 free calls every month — including monitor checks. Notifications (webhook, Slack, email) are always free.
Extract, search, crawl, and scheduled monitor checks all share one rate.