VResource | Resource AI

Context

The team was monitoring hundreds of sources every day, but duplicates and noise were slowing down fact checking. They needed an agent that merges events, removes noise, and explains context quickly.

Solution

We built a Telegram news agent that collects sources, normalizes events, merges duplicates, and delivers personalized digests with threaded discussions.

Bot link: Open Resours AI in Telegram

What is inside

• ingestion from RSS, media outlets, and Telegram channels;
• normalization and event merge logic;
• scoring and decision: push, digest, or skip;
• profile learning from 👍/👎/🔕 reactions and chats;
• on-demand fact checks.

Results

• less noise and repetition;
• faster understanding of context;
• ready to scale sources and users.

Stack

• LLM + embeddings
• PostgreSQL + pgvector
• Redis queues
• Python
• Telegram API

Phases

Discovery → PoC

CTA

We can customize similar agents for your information landscape, sources, and fact-check rules.