Contact person: Business Client

Phone:Show

Email:Show

Location: Kavali - Nellore, India

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I need a lightweight solution that takes a file-based list of URLs, pulls the visible text from each page (articles, landing-page copy - no images or metadata), and lets me query that corpus through an LLM-powered semantic search interface.

1. Input: I will upload a CSV (around 500 URLs per file), with columns:
ID, Type (category), URL



2. Tool should:
Let me upload a new CSV every day (simple web UI is enough). For each URL, crawl and extract only the primary content: Landing page content, main article / body text, Page title. Ignore header, footer, sidebar, menus, ads, etc.

Store: , URL, ID , Type, Title, content



3. Search with LLM / Embeddings (not self-hosted)
I want to use a LLM / embedding API (OpenAI or similar, no self-hosted models). Build a simple search UI where I can:

Enter a string or comma-separated words as input. Example: facebook
Select 1–2 Types from a dropdown (to filter results by the Type column from CSV).



4. The system should:
Use embeddings / semantic search to find relevant URLs where the context matches my input.
It should not only match exact words, but also related concepts.

Example: Input "facebook" -> should also surface pages about Meta, Instagram, Twitter, Mark Zuckerberg, social networks, etc. when contextually relevant.

Search priority:

- First in title (highest priority)
- Then in main content / description



5 Output for each result:
- URL
- Title
- Short snippet (optional but nice)

Label where it matched:

Matched in: Title or
Matched in: Content / Description


Additioanl Note:
- This is a side project, so please keep running costs low (use a reasonably priced AI API).
I don’t care what tech stack you use, as long as:
- It’s a simple web interface (upload + search)
- It’s stable
- You give me basic instructions on how to run it.

NOTE: Please apply only if you have experience with web scraping + using AI APIs (LLMs/embeddings) and can explain in simple terms how your solution will work." (client-provided description)