How do agents connect?
MCP tools or the async job API - both return typed, stable JSON. An agent can call a scraper the way it calls any tool, and the schema does not shift underneath it when a target site redesigns.
Is it fast enough for interactive agents?
~2s median on major sources, ~4s at P95 - built for live tool calls inside a conversation, not overnight batches. Slower, browser-heavy targets are flagged so your agent can set expectations.
Why not let the agent browse the web itself?
Because the open web blocks agents and feeds them HTML soup. A managed scraping layer gives the agent clean, reliable, structured answers - and your tokens go to reasoning, not to parsing markup or fighting anti-bot walls.
How do I get the API documentation and a data sample?
API documentation is public at scrapeguys.com/docs - request format, response schemas, status codes and worked examples. For a sample dataset on your exact targets, book a short scoping call: each deployment is customized to your schema, fields and volumes, and most teams have a sample in hand within a day.
How is the data delivered - do we have to integrate an API?
An API is one option, not a requirement. The same data layer ships as REST JSON APIs with async jobs and webhooks, scheduled exports into Power BI, Looker, Snowflake, BigQuery or Google Sheets, a white-label dashboard your team signs into with zero integration, or MCP tools your AI agents call live. You can mix channels - API for engineering, dashboard for the commercial team - on one deployment.
Is the data collection compliant?
We collect publicly available data only - no login-gated content, no private accounts, no personal data harvesting. Every deployment is reviewed against the target and the use case before it goes live, and enterprise contracts include the data-processing and confidentiality terms your legal team expects. Bring your compliance questions to the scoping call - we answer them in writing.