klbr/n8nworkflows.xyz
1
0
Fork 0
mirror of https://github.com/khoaliber/n8nworkflows.xyz.git synced 2026-04-28 00:29:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

creation

Browse Source
This commit is contained in:
nusquama
2025-11-13 14:34:19 +01:00
parent d850c07552
commit 13cf310e9c
1 changed files with 222 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
workflows/Daily AI News Summaries with Gemini 2.5 Flash to Telegram-6095/readme-6095.md
+222
View File
@@ -0,0 +1,222 @@
Daily AI News Summaries with Gemini 2.5 Flash to Telegram
https://n8nworkflows.xyz/workflows/daily-ai-news-summaries-with-gemini-2-5-flash-to-telegram-6095
# Daily AI News Summaries with Gemini 2.5 Flash to Telegram
---
### 1. Workflow Overview
This workflow automates the daily curation and delivery of AI technology news summaries by integrating real-time news scraping, AI-based analysis, and messaging. It is designed to:
- Continuously monitor and fetch new AI news articles from a specified RSS feed.
- Extract full article content using AI-powered scraping.
- Use Google’s Gemini 2.5 language model (via Langchain integration) to synthesize and generate a concise, well-structured daily briefing.
- Format the briefing for Telegram-compatible HTML and deliver it as a message to a designated Telegram chat.
**Target Use Cases:**
- Automated daily briefings for AI researchers, analysts, or enthusiasts.
- Centralized digest of news from multiple AI news sources.
- Integration into team communication platforms for up-to-date information sharing.
**Logical Blocks:**
- **1.1 Input Reception:** Trigger and fetch new articles from the AI-News RSS feed.
- **1.2 Content Extraction:** Retrieve full article content from each news link using Jina AI.
- **1.3 AI Processing & Report Generation:** Use Gemini 2.5 LLM with a custom prompt to analyze and generate a structured news summary.
- **1.4 Delivery:** Send the formatted briefing as a Telegram message.
---
### 2. Block-by-Block Analysis
#### 1.1 Input Reception
- **Overview:**
This block initiates the workflow by polling the AI-News RSS feed every minute to detect new articles. It outputs metadata including article titles and links.
- **Nodes Involved:**
- AI-News Feed (RSS Feed Read Trigger)
- **Node Details:**
- **AI-News Feed**
- Type: RSS Feed Read Trigger
- Role: Polls the RSS feed at `https://www.artificialintelligence-news.com/feed/` every minute for new items.
- Configuration: Poll mode set to every minute; no filters applied.
- Inputs: None (trigger node)
- Outputs: Emits new article metadata (title, link, publication date, etc.)
- Version Requirements: Standard RSS trigger, no special version dependencies.
- Potential Failures: Network errors, invalid feed URL, feed downtime.
- Sticky Note: Describes this as the workflow start, triggering automatically every minute.
#### 1.2 Content Extraction
- **Overview:**
Processes each new article link from the RSS feed to fetch full article content using Jina AI’s scraping capabilities.
- **Nodes Involved:**
- Read News from AI News Website
- **Node Details:**
- **Read News from AI News Website**
- Type: Jina AI Node
- Role: Scrapes or retrieves the full textual content of the article from the provided URL.
- Configuration: Uses the `link` field from the RSS feed as input URL; no additional options set.
- Inputs: Receives article metadata including the link from AI-News Feed node.
- Outputs: Article full content, cleaned and ready for analysis.
- Credentials: Requires Jina AI API credentials.
- Potential Failures: Link unreachable, content extraction failure, API rate limits or timeouts.
- Sticky Note: Explains it scrapes full article text using Jina AI.
#### 1.3 AI Processing & Report Generation
- **Overview:**
Synthesizes the scraped news articles into a concise, insightful report using the Gemini 2.5 language model via Langchain. The prompt instructs the AI to focus on key news, themes, and formatting compatible with Telegram HTML.
- **Nodes Involved:**
- Gemini 2.5 Flash
- Generate Report
- **Node Details:**
- **Gemini 2.5 Flash**
- Type: Langchain Google Gemini Chat Node (lmChatGoogleGemini)
- Role: Provides AI language model capabilities to process input data and generate text.
- Configuration: Default options; references Google Gemini API credentials for authentication.
- Inputs: Receives prompt and input data from Generate Report node as an AI language model input.
- Outputs: AI-generated text (structured briefing).
- Credentials: Google Palm API credentials required for Gemini access.
- Potential Failures: API quota exceeded, authentication errors, network issues, model timeouts.
- **Generate Report**
- Type: Langchain Chain LLM Node
- Role: Constructs a detailed prompt embedding scraped article content from two sources, instructing the Gemini model to produce a structured daily intelligence briefing with specific formatting rules for Telegram.
- Configuration:
- The prompt includes placeholders for article title, content, and publish date.
- Output restricted to a subset of HTML tags supported by Telegram (<b>, <i>, <u>, <s>, <code>, <pre>, <a>), with lists formed by hyphenated lines.
- The current date is dynamically inserted in British English format.
- Inputs: Receives article content from “Read News from AI News Website” and AI model output from Gemini 2.5 Flash.
- Outputs: Final formatted briefing text for Telegram.
- Version: Version 1.7 of the Langchain Chain LLM node.
- Potential Failures: Expression evaluation errors, prompt formatting issues, incomplete input data, model response inconsistencies.
- Sticky Note: Described as the “brain” of the workflow that uses Gemini to generate the report.
#### 1.4 Delivery
- **Overview:**
Sends the final AI-generated news briefing as an HTML-formatted message to a predefined Telegram chat.
- **Nodes Involved:**
- Send a text message
- **Node Details:**
- **Send a text message**
- Type: Telegram Node
- Role: Sends a text message to a Telegram chat via bot API.
- Configuration:
- Message text set dynamically from the AI-generated report (`={{ $json.text }}`).
- Chat ID set as `CHAT_ID` (to be replaced with actual destination chat).
- Parse mode set to HTML to enable formatted output.
- Inputs: Receives the final text from the Generate Report node.
- Credentials: Requires Telegram API credentials (bot token).
- Potential Failures: Invalid chat ID, bot permissions missing, network errors, message size limits.
- Sticky Note: Final step delivering the briefing to Telegram.
---
### 3. Summary Table
| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
|-------------------------------|----------------------------------|----------------------------------------|-----------------------------|--------------------------|---------------------------------------------------------------------------------------------|
| AI-News Feed | RSS Feed Read Trigger | Triggers on new AI-News RSS articles | None | Read News from AI News Website | Starting point; triggers every minute; fetches new articles from RSS feed. |
| Read News from AI News Website| Jina AI Node | Extracts full article content | AI-News Feed | Generate Report | Scrapes full article content using Jina AI from article link. |
| Gemini 2.5 Flash | Langchain Google Gemini Chat Node| Provides AI model processing | Generate Report (ai_languageModel input) | Generate Report (ai_languageModel output) | Uses Gemini 2.5 LLM for text generation; requires Google Palm API credentials. |
| Generate Report | Langchain Chain LLM Node | Creates structured AI news summary | Read News from AI News Website, Gemini 2.5 Flash | Send a text message | “Brain” node generating final formatted report; uses prompt with Telegram HTML formatting.|
| Send a text message | Telegram Node | Delivers report to Telegram chat | Generate Report | None | Sends AI-generated briefing to Telegram chat with HTML parse mode enabled. |
| Sticky Note | Sticky Note | Documentation and explanations | None | None | Various sticky notes describe each major block and nodes in detail. |
---
### 4. Reproducing the Workflow from Scratch
1. **Create the RSS Feed Trigger Node:**
- Add a node of type **RSS Feed Read Trigger**.
- Configure the feed URL as: `https://www.artificialintelligence-news.com/feed/`.
- Set polling to occur **every minute**.
- Name this node: `AI-News Feed`.
2. **Add Jina AI Node for Article Content Extraction:**
- Add a node of type **Jina AI**.
- Set the **URL** parameter to: `={{ $json.link }}` to dynamically use the article link from RSS feed.
- Leave other options default.
- Connect the output of `AI-News Feed` to this node’s input.
- Provide **Jina AI API Credentials** under the node’s credentials section.
- Name this node: `Read News from AI News Website`.
3. **Set up Gemini 2.5 Flash Node (Google Gemini LLM):**
- Add a node of type **Langchain Google Gemini Chat Node** (`lmChatGoogleGemini`).
- No special parameters needed under options.
- Assign valid **Google Palm API Credentials** to enable Gemini access.
- Name this node: `Gemini 2.5 Flash`.
4. **Add Langchain Chain LLM Node for Report Generation:**
- Add a node of type **Langchain Chain LLM**.
- Configure the prompt to include:
- Instruction as an expert AI analyst synthesizing two news sources.
- Input template embedding scraped article fields: `title`, `content`, `publishedTime`.
- Required output format restricting to Telegram-supported HTML tags only: `<b>, <i>, <u>, <s>, <code>, <pre>, <a>`.
- Use hyphen-prefixed lines for lists instead of `<ul>/<ol>/<li>`.
- Include dynamic date insertion using JavaScript expression:
`{{ new Date().toLocaleDateString('en-GB', { day: 'numeric', month: 'long', year: 'numeric' }) }}`.
- Connect the output of `Read News from AI News Website` as input.
- Set the AI language model input to the `Gemini 2.5 Flash` node.
- Name this node: `Generate Report`.
5. **Add Telegram Node to Send Message:**
- Add a node of type **Telegram**.
- Configure the **Chat ID** to your target Telegram chat (replace `CHAT_ID` with actual chat ID or username).
- Set **Text** field to: `={{ $json.text }}` to send the text output from the previous node.
- Enable **Parse Mode** and set it to **HTML** to ensure formatting is rendered correctly.
- Provide **Telegram Bot API Credentials** to allow sending messages.
- Connect the output of `Generate Report` to this node.
- Name this node: `Send a text message`.
6. **Connect Nodes in Order:**
- `AI-News Feed` → `Read News from AI News Website` → `Generate Report` → `Send a text message`.
- Also connect `Gemini 2.5 Flash` as the AI language model processor input to `Generate Report`.
7. **Set Credentials:**
- Ensure you have valid credentials for:
- Jina AI API
- Google Palm API (for Gemini)
- Telegram Bot API (bot token)
8. **Testing and Deployment:**
- Activate the workflow.
- Verify that new articles trigger the workflow, full content is scraped, the AI generates reports, and the Telegram messages arrive formatted correctly.
---
### 5. General Notes & Resources
| Note Content | Context or Link |
|-------------------------------------------------------------------------------------------------|-----------------------------------------------------------------|
| The workflow uses Telegram-supported HTML tags only for message formatting: `<b>, <i>, <u>, <s>, <code>, <pre>, <a>`. Lists are created with hyphen-prefixed lines. | Important for Telegram message compatibility. |
| Gemini 2.5 Flash node requires Google Palm API credentials with sufficient quota and access. | https://developers.generativeai.google/ |
| Jina AI node scrapes article content from URLs; ensure links are publicly accessible. | https://docs.jina.ai/ |
| Telegram Bot must have permissions to send messages to the target chat, and chat ID must be correct. | https://core.telegram.org/bots/api#sendmessage |
| Workflow triggers every minute; consider API rate limits and polling frequency for scalability. | Can be adjusted for lower frequency if necessary. |
| Sticky notes in the workflow provide detailed explanations of node roles and parameters. | N8N UI sticky notes for workflow documentation and maintenance. |
---
**Disclaimer:**
The text processed and generated by this workflow originates exclusively from publicly available AI news sources and automated AI processing. All operations comply with current content policies and legal guidelines.
---