Build an n8n news radar: Inoreader → Claude/OpenAI → Telegram
Job Details
Description
Summary We are looking for an experienced n8n automation specialist to build a paid prototype of an internal AI-powered news radar for a media team. The goal is to automatically collect news, identify the strongest topics, group duplicate reports, find important details hidden inside articles, and deliver a clear top-news digest to a private Telegram chat. We have already selected the basic architecture: Inoreader folder → self-hosted n8n → Claude/OpenAI API → private Telegram bot/chat This is not a custom SaaS project and should not be developed from scratch. We want to reuse standard n8n nodes, Inoreader functionality and existing integrations wherever possible. How the prototype should work One configurable Inoreader folder will initially contain approximately 15–20 sources, including public Telegram channels, RSS feeds and other sources supported by Inoreader. Every 30 minutes, the workflow should: collect new items from the selected Inoreader folder; avoid processing the same Inoreader item twice; send the available article/post content to Claude or OpenAI using our editable editorial prompt; group reports covering the same real-world event; preserve useful facts, numbers, quotes and unique details from different sources; identify related but separate developments and suggest when they can be presented under one broader editorial angle; detect the strongest and most interesting angle, including important details that may be mentioned inside the article rather than in the original headline; maintain an updated ranking of the strongest current news topics. The AI should create engaging editorial angles and headline ideas, but must not invent facts or present allegations as confirmed facts. Telegram output The Telegram bot should provide the following: 1. /now Show the current top 5 news topics at any moment, based mainly on recent developments from the last several hours. 2. /daily Show approximately the top 10 topics from the last 24 hours. Multiple updates about the same story should be consolidated rather than repeated. 3. Automatic current-news digest Send an updated top-5 digest every 2 hours. The schedule must be editable, and the automatic digest must be possible to disable. 4. Automatic daily digest Send a “Main news of the last 24 hours” digest once per day. The delivery time must be configurable, and the daily digest must be possible to disable. 5. Breaking-topic alert Send an immediate alert only when a genuinely strong new topic appears. The same story should not be repeatedly sent unless there is a significant new development. Format of each topic Each selected topic should include: one strong primary headline; two alternative headline options; a concise factual summary; important unique details collected from different sources; a note when several related developments may be combined; clickable links to every original article or Telegram post used. Long URLs should be hidden inside short source names, for example: Sources: AP, Reuters, original Telegram post The AI should not generate or rewrite URLs. n8n should preserve the original URLs received from Inoreader and insert them into the Telegram message. Prompt management The editorial prompt may be long and may include multiple examples of: original news items; preferred editorial angles; stronger but factually accurate headlines; unacceptable exaggerations; examples of important details hidden inside longer articles. We must be able to edit the prompt at any time without changing workflow code or contacting the developer. Prompt changes should apply to the next workflow run. Source management Sources will be managed inside the Inoreader folder. Adding, removing or moving a source inside that folder should not require any modification of the n8n workflow. The first prototype will use one working Inoreader folder and one editorial profile. Future editorial profiles The workflow should be structured so that additional editorial profiles can later be added through configuration rather than rebuilding the system. Each future profile should be able to use its own: Inoreader folder; editorial prompt and examples; output language; Telegram destination; digest schedule; enabled/disabled status. Possible future profiles include Russian, Belarusian, English and Polish editorial teams. Building all profiles is not required in the prototype. The prototype should simply avoid hardcoding the entire system around one profile. Deliverables The contractor should provide: a working workflow inside our n8n account/server; connection to our Inoreader folder; connection to our Claude or OpenAI API account; private Telegram bot or output chat; /now command; /daily command; automatic top-5 digest every 2 hours; configurable automatic daily digest; breaking-topic alerts; editable editorial prompt; prevention of duplicate item processing; clickable links to all original sources; exported n8n workflow JSON; short written setup instructions; short video walkthrough. The workflow must use our own accounts, credentials and infrastructure. Prototype limitations Full-text extraction from external websites is not required for the first prototype. The workflow should initially use the content provided by Inoreader. Additional full-text extraction may be discussed later only for specific sources where Inoreader content is insufficient. Not required The prototype does not require: a custom website or dashboard; CRM or approval workflow; take/reject/postpone buttons; automatic public publishing; automatic analysis of all images and videos; a direct Telegram userbot; complete YouTube script generation; multiple fully configured editorial teams; a custom backend or SaaS platform. Budget and timeline Budget: $250 fixed price Expected delivery: 5–7 working days. This is a paid prototype. If the prototype works well, we plan to continue with additional paid milestones for: expansion to more sources; additional editorial profiles; improved error handling and recovery; optional media analysis; more advanced editorial outputs. Please include a screenshot or short video of a relevant n8n workflow you have built. Please do not propose development from scratch or a custom SaaS platform.
Skills
Artificial Intelligence API Integration n8n Telegram API RSSNotification History
| Channel | Type | Status | Sent | Error |
|---|---|---|---|---|
| telegram | pre_ai_job_alert | sent | Sat, Jul 18, 2026 3:28 PM | - |
User Actions
| Action | Acted at |
|---|---|
| No actions. | |