We built an AI news pipeline that collects articles from 150+ RSS feeds, groups duplicate stories, scores relevance and writes short Hindi and English summaries. Editors approve every story before it goes out on web and mobile.
Industry: SaaS & TechnologyService: AI & AutomationUpdated
Overview
The challenge
News arrived all day from more than 150 RSS feeds. The same story was often reported by many sources at once, and editors had to read through long articles to find what was new, important and relevant to their national, state and city readers.
Each chosen story then had to be rewritten as a short, readable item with a clear headline, in both Hindi and English. Doing all of this by hand limited how much news the team could cover and how quickly it reached readers.
The same story from many sources filled the queue and made genuinely new updates easy to miss.
Turning long articles into short Hindi and English items took editors a lot of time.
With so many articles arriving, it was hard to see which national, state and city stories mattered most.
The solution
n8n workflows pull articles from more than 150 RSS feeds, clean them up and pass each one through the AI steps before it reaches the editors.
AI finds articles that cover the same underlying story, groups them together and keeps genuinely new updates separate. Each story gets a relevance score so editors see the most important news first.
For each story, AI writes a short summary and a clear headline in Hindi and English. Editors review, edit and approve every item in the administration panel before it is published to web and mobile.
Before and after
Before
Reading every feed
Editors scanned many sources and often read the same story several times.
After
One grouped, ranked queue
Duplicate coverage is grouped and stories arrive ranked by relevance.
Before
Rewritten by hand
Each long article was condensed and written up in two languages manually.
After
Drafted by AI
Short Hindi and English summaries and headlines are ready for editors to review.
Before
Slow to reach readers
Manual sorting and rewriting limited how much news the team could publish.
After
Approve and publish
Editors check, adjust and approve each item, then publish it to web and mobile.
Key features
AI turns full-length news articles into short, readable summaries in Hindi and English, each with a clear, human-friendly headline.
Business impact
Less rewriting for editors
Semantic matching spots articles about the same underlying story across sources, groups the coverage and helps editors pick the best article while keeping genuinely new updates.
Business impact
A cleaner review queue
Each story is scored for importance and relevance, so editors can prioritise national, state and city news for review and publication.
Business impact
Clearer editorial priorities
Results
150+ Feeds in One Queue
News from 150+ RSS feeds arrives in a single queue, with duplicate coverage already grouped.
Before: Many FeedsAfter: One Queue
Less Manual Rewriting
Editors start from AI-written Hindi and English summaries instead of condensing every article by hand.
Before: Written by HandAfter: AI Drafted
Every Story Approved by an Editor
Every item is reviewed and approved by an editor before it is published to web and mobile.
Before: Manual SortingAfter: Review and Approve
Faster delivery
Building the pipeline from separate n8n steps lets the team add new feeds or change an AI step without rebuilding the rest of the workflow.
n8n workflows collect and clean articles from 150+ RSS feeds on a schedule, so new stories enter one queue.
Separate steps for summarisation, duplicate detection and relevance scoring, which can each be tuned on their own.
Editors see grouped, ranked stories with their Hindi and English drafts, and approve, edit or reject each one.
Approved stories go out to the web and mobile platforms from the same panel.
Trust and control
Nothing is published automatically: an editor reviews, edits and approves every story first.
HUMAN APPROVALEach summary stays linked to its original articles, so editors can check the facts against the source.
TRACEABLEDuplicate grouping preserves genuinely new developments, so follow-up stories are not lost among repeats.
ACCURATEFeeds, scoring and summary steps can be changed as the platform's coverage and priorities change.
ADAPTABLEThe pipeline collects articles from more than 150 RSS feeds using n8n workflows. AI then groups duplicate coverage of the same story, scores each story for relevance and writes short summaries with headlines in Hindi and English. Editors review and approve every item in an administration panel before it is published to web and mobile.
Semantic duplicate detection compares articles by meaning rather than exact wording, so it finds coverage of the same underlying story across sources. These articles are grouped together, editors pick the most suitable one, and genuinely new updates on a developing story are kept separate so they are not lost among repeated reports.
No. The AI prepares summaries, headlines, duplicate groups and relevance scores, but nothing is published automatically. An editor reviews each story in the administration panel, can edit the Hindi and English text, and approves it before it goes out to the web and mobile platforms, so the newsroom keeps full editorial control.
If your editors read the same story from ten sources and rewrite every item by hand, we can build a pipeline that groups, ranks and drafts it for them.