Getting started
Create a workspace, connect a client, and inspect source-backed results.
1. Sign in and create a workspace
Create an account, verify your email, and name your workspace. Existing users can sign in. Invitations grant access only to the addressed account.
2. Choose a workflow
The console groups nineteen tools into discovery, collection, operations, and monitoring. Start with Scrape for one public URL or Map to discover scoped URLs.
Choose bounded inputs. A discovered URL is not proof that its page was fetched. Partial captures and unavailable providers are reported explicitly.
3. Connect your application or agent
- For REST, create a scoped API key. Copy its one-time secret into your application's private configuration. Never put it in browser source or commit it to Git.
- For MCP, use the endpoint shown in MCP connections. A compatible remote MCP client can open OAuth sign-in and request your consent for one workspace and selected tools.
- Download the hosted Postman collection or inspect the API reference.
4. Inspect the result and evidence
Jobs retain operation status, errors, and result references. Artifacts expose authorized metadata and bounded, integrity-checked downloads. Datasets organize captured versions with operation lineage.
Use a new idempotency key for each new execution. Retain that key and its original arguments if a response is uncertain. Check job history before resubmitting.
Current execution boundary
Account access, workspace permissions, scoped keys, and MCP consent are deployed. Hosted capture requires the separately configured engine and its service identity. Until connected, execution returns an explicit unavailable error. It does not fabricate results.
Developers working on the engine can use the separate local setup guide. Local loopback servers must remain private.