← All projects

AI & AUTOMATION / MEDIA AUTOMATION

Telegram Autoposter.

An autonomous pipeline that collects, summarizes, enriches, and posts news to Telegram, with an LLM handling relevance filtering.

SECTOR

Media Automation

LOCATION

Eastern Europe

FOCUS

Blog / Content Automation

DELIVERY

Founder’s freelance project

Concept illustration of news articles passing through a processing ring into a paper plane
Conceptual project illustration. Client interface and data are not shown.

01 / THE CHALLENGE

The problem worth solving.

Keeping a news channel fresh meant manually finding, summarizing, and posting stories.

What the system needed to do

  • Handle unstable and inconsistently structured RSS feeds
  • Use an LLM for relevance and quality filtering, not just formatting
  • Leave room to extend to future platforms such as Pinterest

The real-world constraints

  • LLM cost and rate limits at scale
  • Fragile, inconsistent RSS structures across sources

02 / THE SOLUTION

How it came together.

We deployed a resilient, VM-based autoposter running async pipelines with configurable schedules and automatic recovery from source failures.

Normalize RSS articles and track their source identifiers.

Explore the architecture. This is an explanatory diagram, not a live production system.

03 / THE OUTCOME

What the work delivered.

  • Low-maintenance operation with automatic recovery
  • Configurable posting schedule with no manual triggering

YOUR NEXT CHAPTER

Something great
starts with a conversation.

Bring one repetitive process. We’ll discuss the time it takes, what could be automated, and the next step that fits your business.

30 minutes. One workflow. No obligation.