MIYAIP · Web Data
Choose how to collect web data
Start with the result you need. Submit a structured crawler task, let an AI client work in a browser, or connect your own scraper through a proxy.
Choose what MIYAIP runs—and what you control
| Your task | Start here | You provide | What happens next |
|---|---|---|---|
| Collect data through a supported interface | Crawler API | An interface and its required parameters | Submit a task, keep its task ID, and retrieve status, results and credit usage |
| Navigate, click or extract data in a browser with an AI client | Browser MCP — Beta | A compatible MCP client, an MCP key and a task | The agent operates a JavaScript browser and receives observations and task artifacts |
| Run your own HTTP client, scraper or browser automation | Proxy services | Your code, target and parsing logic | Configure a proxy endpoint; your application controls requests, retries and output |
The endpoint and credentials differ between these routes. Proxy usernames, console login tokens and MCP keys serve different connections.
Open the service that fits your workflow
Crawler API
Run asynchronous data collection tasks through the available interface catalog. Choose a supported Google, Reddit, TikTok or website interface, enter its parameters, and retrieve the result by task ID. Output fields and credit usage depend on the selected interface.
Browser MCP
BetaConnect an MCP-compatible AI client to a real JavaScript browser. Use it for tasks that require page interaction, screenshots or structured extraction across browser steps. Create an MCP key and follow the configuration for your client.
Proxy services
Keep control of your crawler or browser automation. Choose a fixed or rotating exit and the billing model that fits your workload, then add the supplied connection fields to your application.
Google search results
Looking specifically for search results? Open the SERP API product page. For an asynchronous workflow using the Crawler API catalog, follow the Google Search interface documentation.
From a service choice to a working request
Choose the execution model.
Decide whether you need a crawler task, an interactive browser or a proxy connection for your own application.
Open the matching setup guide.
Check the interface, authentication method and input fields for that service.
Run one small task.
For Crawler API, retain the task ID and inspect the result. For Browser MCP, confirm the tools are available in your client. For a proxy, verify the connection and observed exit IP.
Check the output before scaling.
Review the actual returned fields, errors, timing and applicable usage charges against the task you intended to complete.
Frequently asked questions
Do I need a proxy if I use Crawler API?
The Crawler API workflow uses its own interface and task parameters. A separate proxy connection is the route for applications that send and process their own requests. Follow the selected service's setup instead of adding proxy credentials to an API request.
How is Browser MCP different from a crawler task?
Browser MCP lets an AI client interact with a browser over multiple steps. Crawler API accepts an interface invocation and returns a task ID that your application uses to retrieve status and results.
Can I use one credential for all services?
No. A proxy endpoint may use a proxy username and password. Crawler API uses the console login token described in its documentation. Browser MCP uses a separately created MCP key.
Does every crawler interface return the same JSON fields?
No. Inputs and outputs depend on the interface. Read its schema and inspect a small result before connecting it to a downstream process.
Where should I compare proxy costs?
Use proxy pricing to compare the available products and billing units. API and browser task usage follow their own product documentation.
Compare proxy pricingChoose a route and complete your first task
Open the relevant setup guide, connect the service, and inspect the result before increasing your workload.