A: Hey amanybusiness,

Yes, absolutely! AgenticFlow is designed for exactly this type of task. Here’s a guide on how you can achieve it:

1. How to Create the Workflow:

You'll use our Workflow Builder for this. The basic steps would be:

Step 1: Get the Article Content:
Add the Web Scraping node (web_scraping).
In its input, provide the specific article URL you want to scrape.
Set the scraping_type to Text (this usually works best for extracting content for AI analysis).
Step 2: Extract Numeric Data with AI:
Add an LLM node (Large Language Model, like our built-in Gemini-Flash or 4o-mini, or you can use your own key - BYOK - for others like GPT-4/Claude).
Connect the scraped_content output from the Web Scraping node to the input of the LLM node.
Crucially, provide a clear prompt telling the AI exactly what numeric data to find. Example prompt:
From the following article text: {{web_scraping.scraped_content}}

Please extract the following numeric values related to [mention the aspects, e.g., company revenue, user count, market share]:
- Revenue Figure:
- Number of Users:
- Market Share Percentage:

Return ONLY the numbers found for these aspects. If a value isn't mentioned, state "Not Found".
content_copy
download
Use code with caution.
(Adapt the aspects you want to extract based on the article)
Step 3: Get the Output:
The output of the LLM node (content) will contain the extracted numeric data as requested in your prompt. You can then use this data in subsequent steps (e.g., save it to a table, send it in an email).
2. Making Connections:

AgenticFlow's Built-in Nodes: For the core Web Scraping and built-in LLM nodes, you don't need to set up a separate "connection" in the traditional sense. They use the platform's underlying capabilities and your account's credits.
External Tools (Optional): If you wanted to use a more advanced external scraper (like Apify via its MCP: https://agenticflow.ai/mcp/apify) or a specific external LLM with your own API key (BYOK), then you would need to set up a Connection or provide the key within that specific node's configuration.
3. Costs:

No Extra Cost for Connections: Setting up connections (like providing an API key for an external service) does not cost extra.
Credit Usage: The purchase of AgenticFlow (your Tier 2 plan) gives you a monthly pool of credits. Running the workflow will consume credits:
Each step (like Web Scraping, LLM) has a Fixed Cost (4 credits per step on Tier 1/2).
Steps using AgenticFlow's built-in LLMs also have a Variable Cost based on the amount of text processed (input/output tokens).
If you use BYOK for an LLM, you avoid the variable credit cost for that specific LLM call (you pay the provider directly), but the step still uses the fixed 4 credits.
This workflow (Scrape -> LLM Extract) would likely consume around 8 fixed credits + variable credits for the LLM extraction per run.
4. Step-by-Step Instructions/Tutorials:

General Workflow Building: Our documentation covers the basics of creating workflows: [Link to Workflow Docs - if available, otherwise omit or point to general docs]
Template Example: A good starting template might be the "Company Research w/ website" (https://agenticflow.ai/app/explore/workflow/d4ab367e-3a6c-4156-94a0-f7266722d2ff). You can clone this and modify the LLM prompt significantly to focus on extracting numeric data instead of just summarizing.
Specific Help: If you get stuck, feel free to post specific questions in our community forum (https://community.agenticflow.ai/), and we can guide you further!
Let me know if you have more questions as you start building!