I designed and built an AI Web Research Agent using Python, Streamlit, OpenRouter-compatible APIs...I designed and built an AI Web Research Agent using Python, Streamlit, OpenRouter-compatible APIs...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
I designed and built an AI Web Research Agent using Python, Streamlit, OpenRouter-compatible APIs, DuckDuckGo search, Requests, and Trafilatura. The workflow creates targeted search queries, gathers and ranks results, reads available webpages, preserves source URLs, distinguishes fully fetched pages from search-result-only sources, and generates a structured Markdown report. I also created the interface, live activity feed, quick demo prompts, source-quality labels, verification-limit section, and downloadable report experience. The output is designed as a research draft that remains subject to human review.
Post image
Post image
Post image
Post image
Eric's avatar
The fetched-versus-search-only label is the part I'd keep visible in every section. If I can tell what the agent actually read and what it only saw in search results, I know which claims need another click before I trust them.
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started