Field manual · Unit WR-01 · Component inventory

Web Scraping Software Built Around an AI Agent, Not Selectors

WebRobot is web scraping software where an AI agent does the work: it reads pages like a human, extracts the fields you describe in plain English, monitors sites on a schedule and delivers clean data to Sheets, Slack and your API. Here is every module in the unit.

Run the robot

01 / THE AGENT FIG. 1 · EXTRACTION CORE

An extraction agent that survives redesigns

The core of the unit: no selectors, no recorded clicks, no re-training after every site update.

MODULE A1

Plain-English setup

Describe the data in one sentence and paste a URL. The agent builds the field schema and browsing plan itself. Setup for a typical catalog robot takes under ten minutes, and there is no code anywhere, as on any no code web scraper.

input: 1 sentence · setup: <10 min · code: none

MODULE A2

Human-like browsing

A real browser session clicks, scrolls, paginates, signs in and fills forms with human pacing. Everything a person can reach in a browser, the robot can extract, which makes it a full browser automation tool.

actions: click · type · scroll · login · paginate

MODULE A3

Self-healing extraction

When a site redesigns, the agent re-reads the page and re-maps your fields automatically. Fleet-wide self-heal rate is 98.6% with an average repair of 41 seconds. Broken runs pause and alert instead of shipping wrong rows.

self-heal: 98.6% · avg repair: 41s

MODULE A4

Schema validation + dedupe

Every row is type-checked against your schema, deduplicated across pages and runs, and flagged when values look wrong. You get analysis-ready tables, not HTML soup.

checks: type · format · required · dedupe: cross-run

02 / SCHEDULING + MONITORING FIG. 2 · TIMING SUBSYSTEM

Schedules, monitors and change alerts

Data that refreshes itself. Set the cadence once and the fleet keeps it.

MODULE B1

Daily to real-time schedules

Run robots once, daily, hourly or as real-time monitors depending on plan. Fresh rows land before your morning coffee, every day, without anyone pressing a button.

cadence: once · daily · hourly · real-time

MODULE B2

Change detection + alerts

Monitors watch prices, stock, listings or whole pages and fire a Slack message or webhook the moment something moves. Ideal for website change monitoring across competitor sets.

triggers: value change · new row · page diff

MODULE B3

Run logs + health dashboard

Every run is logged step by step: pages visited, rows extracted, fields re-mapped, warnings raised. When something needs your attention, you see exactly what and why.

log detail: per-action · retention: full history

03 / DELIVERY + INTEGRATIONS FIG. 3 · OUTPUT MANIFOLD

Delivery to Sheets, Slack, webhooks and a full API

Extraction is half the job. The other half is landing data where your team already works.

MODULE C1

Google Sheets that refresh themselves

Connect a sheet once and the robot keeps it current on every run: new rows appended or replaced, your formulas untouched. CSV and Excel exports are one click.

formats: sheets · csv · xlsx

MODULE C2

REST API + webhooks

Engineers get clean JSON from a documented REST endpoint plus webhooks on every run or change event. WebRobot doubles as a web scraping api.

output: json · auth: bearer token · events: run · change

MODULE C3

Slack + Zapier

Change alerts and run summaries post straight to a channel, and Zapier connects extracted rows to 6,000+ downstream apps: CRMs, databases, email tools.

channels: slack · zapier · webhook

04 / TEAM + ENTERPRISE FIG. 4 · COMMAND STRUCTURE

Seats, roles and enterprise controls

From a single analyst to a procurement-reviewed deployment.

MODULE D1

Team seats + shared robots

Plans include 1 to 15 seats (unlimited on Enterprise). Everyone on the team sees the same robots, schedules and run history, so data work stops living in one person's browser.

seats: 1 / 5 / 15 / unlimited

MODULE D2

SSO, roles and audit log

Enterprise adds SSO/SAML, role-based permissions and a full audit log of who changed which robot and when. Security posture is documented on the security page.

enterprise: sso/saml · rbac · audit log

MODULE D3

Compliance-aware crawling

Robots honor robots.txt, apply rate limits and avoid personal data by default. Data is encrypted in transit (TLS 1.3) and at rest (AES-256), with a DPA and 99.9% SLA on Enterprise.

tls 1.3 · aes-256 · sla: 99.9%

Every module above is included by plan tier, flat monthly pricing and no credit math. See exactly what each tier includes on the WebRobot pricing page.