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
2025-11-25 10:37:42 +01:00
parent c33ddc8a48
commit 267fe0b563
1 changed files with 290 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
workflows/Analyze Images & PDFs from Google Drive with Gemini 3 AI-11038/readme-11038.md
+290
View File
@@ -0,0 +1,290 @@
Analyze Images & PDFs from Google Drive with Gemini 3 AI
https://n8nworkflows.xyz/workflows/analyze-images---pdfs-from-google-drive-with-gemini-3-ai-11038
# Analyze Images & PDFs from Google Drive with Gemini 3 AI
### 1. Workflow Overview
This workflow automates the analysis and summarization of newly uploaded images and PDF financial reports in a specified Google Drive folder using Google Gemini 3 AI models. It is designed for finance professionals, analysts, or anyone needing quick AI-driven insights from visual financial data and documents.
**Logical Blocks:**
- **1.1 Input Reception:**
Watches a specific Google Drive folder for new file uploads (images or PDFs).
- **1.2 File Type Filtering:**
Filters incoming files by MIME type to separate image files from PDFs.
- **1.3 Downloading Files:**
Downloads the actual content of the detected files via HTTP requests.
- **1.4 Image Analysis Branch:**
Sends downloaded images to Google Gemini 3 Vision model for analysis, then processes the output through an AI agent for summarization and key insights extraction.
- **1.5 PDF Extraction and Analysis Branch:**
Extracts text content from PDFs, sends extracted text to Gemini 3 via OpenRouter Chat model, and then passes it through an AI agent for summarization and key insights.
- **1.6 User Guidance and Documentation:**
Sticky Notes provide instructional content, visualization of workflow sections, and links to external resources.
---
### 2. Block-by-Block Analysis
#### 2.1 Input Reception
- **Overview:**
Detects new files created in a specific Google Drive folder, triggering the workflow.
- **Nodes Involved:**
- Google Drive Trigger
- **Node Details:**
- **Google Drive Trigger**
- Type: Trigger node, listens for Google Drive events
- Configuration: Watches a specified folder for `fileCreated` event, polling every minute
- Key expressions: Folder ID needs to be replaced with target folder ID
- Input: None (event-driven)
- Output: JSON metadata of the new file(s)
- Failure modes: OAuth credential issues, folder permissions, API rate limits
- Notes: Folder ID must be customized to user’s folder; credential must have read access.
#### 2.2 File Type Filtering
- **Overview:**
Separates files based on MIME type to route images and PDFs to appropriate processing branches.
- **Nodes Involved:**
- Filter by Type
- **Node Details:**
- **Filter by Type (If node)**
- Type: Conditional logic node
- Configuration: Checks if MIME type equals `image/png` or `image/webp` to identify images
- Input: File metadata from Google Drive Trigger
- Output: Two outputs —
- Output 1: If image MIME type matches
- Output 2: Else, assumed PDF (no explicit condition, but falls through)
- Failure modes: Missing or unexpected MIME types; files of unsupported formats are ignored
#### 2.3 Downloading Files
- **Overview:**
Downloads the actual files from Google Drive using their web content links for further processing.
- **Nodes Involved:**
- Download Image
- Download PDF
- **Node Details:**
- **Download Image**
- Type: HTTP Request node
- Configuration: Downloads image file from URL extracted from file metadata (`webContentLink`)
- Input: From Filter by Type (image path)
- Output: Binary image data
- Failure modes: URL invalid, permissions denied, network failures
- **Download PDF**
- Type: HTTP Request node
- Configuration: Downloads PDF file using `webContentLink` URL
- Input: From Filter by Type (PDF path)
- Output: Binary PDF data
- Failure modes: Same as Download Image
#### 2.4 Image Analysis Branch
- **Overview:**
Processes downloaded images through Google Gemini 3 Vision AI to analyze and summarize visual content (charts, diagrams).
- **Nodes Involved:**
- Analyze an image
- Analyzer Agent
- **Node Details:**
- **Analyze an image**
- Type: LangChain Google Gemini node (Vision model)
- Configuration:
- Model: `models/gemini-3-pro-preview`
- Operation: Analyze image input (binary)
- Text prompt: "Summarize this report/image."
- Input: Binary image from Download Image
- Output: JSON analysis results including textual and visual content interpretation
- Failure modes: API credential issues, model availability, unsupported image formats
- **Analyzer Agent**
- Type: LangChain Agent node
- Configuration:
- Input text extracted from Analyze an image output (`$json.content.parts[0].text`)
- System message instructs to act as a financial analyst assistant, summarizing and highlighting top 3 key findings
- Input: JSON text from Analyze an image node
- Output: Summarized findings
- Failure modes: Parsing errors, AI model latency or failures
#### 2.5 PDF Extraction and Analysis Branch
- **Overview:**
Extracts text from PDF files, processes it through Gemini 3 language model via OpenRouter, and summarizes content with an AI agent.
- **Nodes Involved:**
- Extract from File
- OpenRouter Chat Model1
- Analyzer Agent (PDF)
- **Node Details:**
- **Extract from File**
- Type: ExtractFromFile node
- Configuration: Extracts text content from PDF binary
- Input: Binary PDF from Download PDF
- Output: Extracted text content (JSON)
- Failure modes: PDFs without selectable text (scanned images only), extraction failures
- **OpenRouter Chat Model1**
- Type: LangChain OpenRouter Chat Model node
- Configuration:
- Model: `google/gemini-3-pro-preview`
- Uses extracted text as input
- Input: Extracted text from Extract from File
- Output: Processed text ready for AI agent
- Failure modes: API limits, connectivity, credential issues
- **Analyzer Agent (PDF)**
- Type: LangChain Agent node
- Configuration:
- Input text from OpenRouter Chat Model1 output (`$json.content.parts[0].text`)
- System message: same financial analyst assistant prompt as image branch
- Input: Text processed by OpenRouter Chat Model1
- Output: Summarized report with key insights
- Failure modes: Text parsing issues, AI model errors
#### 2.6 User Guidance and Documentation
- **Overview:**
Sticky Notes provide context, section labeling, and external resource links for users.
- **Nodes Involved:**
- Sticky Note (Image Analyzer)
- Sticky Note1 (PDF Extractor & Analyzer)
- Sticky Note2 (Drive Trigger + Filter by type)
- Sticky Note3 (Full workflow description and setup instructions)
- Sticky Note4 (YouTube tutorial link)
- **Node Details:**
- **Sticky Notes**
- Type: Visual annotation nodes
- Configuration: Provide descriptions, instructions, workflow segmentation, and external video tutorial links
- Input/Output: None
- Purpose: Aid user understanding and onboarding
---
### 3. Summary Table
| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
|---------------------|---------------------------------|------------------------------------|------------------------|-----------------------------|---------------------------------------------------------------------------------------------------------------------------------------|
| Google Drive Trigger | Google Drive Trigger | Detect new files in Drive folder | — | Filter by Type | Drive Trigger + Filter by type |
| Filter by Type | If | Route files by MIME type | Google Drive Trigger | Download Image, Download PDF | Drive Trigger + Filter by type |
| Download Image | HTTP Request | Download image file from URL | Filter by Type (image) | Analyze an image | Image Analyzer |
| Analyze an image | LangChain Google Gemini (Vision)| Analyze image content with Gemini 3| Download Image | Analyzer Agent | Image Analyzer |
| Analyzer Agent | LangChain Agent | Summarize image analysis output | Analyze an image | — | Image Analyzer |
| Download PDF | HTTP Request | Download PDF file from URL | Filter by Type (PDF) | Extract from File | PDF Extractor & Analyzer |
| Extract from File | ExtractFromFile | Extract text from downloaded PDF | Download PDF | Analyzer Agent (PDF) | PDF Extractor & Analyzer |
| OpenRouter Chat Model1 | LangChain OpenRouter Chat Model| Process extracted PDF text with Gemini 3 | Extract from File | Analyzer Agent (PDF) | PDF Extractor & Analyzer |
| Analyzer Agent (PDF)| LangChain Agent | Summarize PDF text analysis output | OpenRouter Chat Model1 | — | PDF Extractor & Analyzer |
| Sticky Note | Sticky Note | Visual annotation for Image branch | — | — | Image Analyzer |
| Sticky Note1 | Sticky Note | Visual annotation for PDF branch | — | — | PDF Extractor & Analyzer |
| Sticky Note2 | Sticky Note | Visual annotation for Trigger/filter| — | — | Drive Trigger + Filter by type |
| Sticky Note3 | Sticky Note | Full workflow description and guide| — | — | Gemini 3 Image & PDF Extractor overview and setup guide with troubleshooting and video tutorial link |
| Sticky Note4 | Sticky Note | YouTube tutorial link | — | — | Link to video tutorial: https://www.youtube.com/watch?v=UuWYT_uXiw0 |
---
### 4. Reproducing the Workflow from Scratch
1. **Create Google Drive Trigger node:**
- Type: Google Drive Trigger
- Event: `fileCreated`
- Folder to watch: Replace with your Google Drive folder ID
- Polling: Every minute
- Credential: Select your Google Drive OAuth credential with read access to the folder
2. **Add an If node (Filter by Type):**
- Conditions:
- Check if MIME type equals `image/png` OR `image/webp` (case-sensitive, strict)
- Input: Connect from Google Drive Trigger
- Outputs:
- True (image path)
- False (assumed PDF path)
3. **Add HTTP Request node (Download Image):**
- Input: Connect from `True` output of Filter by Type
- Method: GET (default)
- URL: Expression `{{$json.webContentLink}}`
- Output: Binary data of image
4. **Add LangChain Google Gemini node (Analyze an image):**
- Input: Connect from Download Image
- Model ID: `models/gemini-3-pro-preview`
- Resource: Image
- Operation: Analyze
- Input Type: Binary
- Text prompt: "Summarize this report/image."
5. **Add LangChain Agent node (Analyzer Agent):**
- Input: Connect from Analyze an image node
- Text input: Expression `{{$json.content.parts[0].text}}`
- System message:
_"You are a helpful financial analyst assistant. You will receive an input of a description of some financial charts extracted from a report. Summarize the chart/information contained in the chart, highlight the top 3 key findings."_
- Prompt type: Define
6. **Add HTTP Request node (Download PDF):**
- Input: Connect from `False` output of Filter by Type
- Method: GET
- URL: Expression `{{$json.webContentLink}}`
- Output: Binary PDF data
7. **Add Extract From File node:**
- Input: Connect from Download PDF
- Operation: PDF text extraction
8. **Add LangChain OpenRouter Chat Model node:**
- Input: Connect from Extract From File
- Model: `google/gemini-3-pro-preview`
- Use default options
9. **Add LangChain Agent node (Analyzer Agent (PDF)):**
- Input: Connect from OpenRouter Chat Model node
- Text input: Expression `{{$json.content.parts[0].text}}`
- System message: Same as Analyzer Agent node for image branch
- Prompt type: Define
10. **Add Sticky Note nodes for user guidance:**
- Position and content as desired, replicating the notes:
- Drive Trigger + Filter by type
- Image Analyzer
- PDF Extractor & Analyzer
- Full workflow description and troubleshooting guide
- YouTube tutorial link
---
### 5. General Notes & Resources
| Note Content | Context or Link |
|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------|
| Gemini 3 Image & PDF Extractor (Google Drive → Gemini 3 → Summary) Automatically summarize newly uploaded images or PDF reports using Google Gemini 3, triggered directly from a Google Drive folder. Perfect for fast AI-powered analysis of financial reports, charts, screenshots, or scanned documents. Includes setup instructions, troubleshooting tips, and a video tutorial link. | Full workflow description in Sticky Note3 node. |
| Video walkthrough of workflow setup and execution: https://www.youtube.com/watch?v=UuWYT_uXiw0 | Sticky Note4 contains the link and reference to the video tutorial. |
| Troubleshooting tips: Ensure Google Drive OAuth credential has read access; verify model IDs for Gemini 3; PDFs must contain selectable text for extraction to succeed; API limits and connectivity issues may affect AI nodes. | Mentioned in Sticky Note3. |
| Supported image MIME types: `image/png`, `image/webp`. PDFs are detected by exclusion and processed separately. | From Filter by Type node and workflow logic. |
---
**Disclaimer:**
The provided content is extracted from an automated n8n workflow using publicly available, legal data and complies with all relevant content policies. No illegal, offensive, or protected data is processed.