diff --git a/docs/tools/tool/agentql.mdx b/docs/tools/tool/agentql.mdx index f5a0389..9cd70f3 100644 --- a/docs/tools/tool/agentql.mdx +++ b/docs/tools/tool/agentql.mdx @@ -9,10 +9,7 @@ Create your own AgentQL API key [here](https://dev.agentql.com). ## Description -AgentQL Web Loader is powered by AgentQL, an AI-powered query language for scraping web sites and automating workflows. -If you want to extract data in a precise format, use [AgentQL query language](https://docs.agentql.com/agentql-query) under the `query` field to pinpoint data on any web page, including authenticated and dynamically generated content. -Users can define structured data output and apply transforms within queries. -You could also directly describe the data you want to extract under the `prompt` field to let AgentQL Web Loader automatically generate the format and data. +[AgentQL](https://agentql.com) provides structured data extraction from any web page using an [AgentQL query](https://docs.agentql.com/concepts/query-language) or a Natural Language prompt. AgentQL can be used across multiple languages and web pages without breaking over time and change. ## Installation @@ -30,15 +27,15 @@ AGENTQL_API_KEY=... The following parameters can be used to customize the `AgentQL Web Loader`'s behavior: -| Argument | Type | Description | -| :--------- | :------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------ | -| **url** | `string` | URL of the website to scrape from. | -| **query** | `string` | _Optional_. AgentQL query to scrape the URL. Please visit [AgentQL Query Language Introduction](https://docs.agentql.com/agentql-query) for more information. | -| **prompt** | `string` | _Optional_. Natural Language description of the data you want to scrape. Either `query` or `prompt` is required. | - -If you want to hack our tool, feel free to do so by modifying `src/tools/agentql_tool.py` and reference our documentation for [AgentQL REST API](https://docs.agentql.com/rest-api/api-reference). +| Argument | Type | Description | +| :-------------------------- | :-------- | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| **url** | `string` | The URL of the web page you want to query. | +| **query** | `string` | _Optional_. The AgentQL query to execute. Learn more about [how to write an AgentQL query in the docs](https://docs.agentql.com/agentql-query). | +| **prompt** | `string` | _Optional_. A Natural Language description of the data to query the page for. AgentQL will infer the data’s structure from your prompt. **Note: You must define either a `query` or a `prompt` to use AgentQL.** | +| **is_stealth_mode_enabled** | `boolean` | Whether to enable experimental anti-bot evasion strategies. This feature may not work for all websites at all times. Data extraction may take longer to complete with this mode enabled. **Defaults to `False`.** | ## Examples - [Research Assistant](https://github.com/AgentOps-AI/AgentStack/tree/main/examples/research_assistant) - [Sentiment Analyzer](https://github.com/AgentOps-AI/AgentStack/tree/main/examples/sentiment_analyser) +- [Market Monitoring](https://github.com/AgentOps-AI/AgentStack/tree/main/examples/market_monitoring)