Product Overview

What is OctoReport

OctoReport is an intelligent content aggregation and report generation platform that helps you:

• Automatically collect content from multiple data sources
• Generate structured reports using AI
• Quickly explore your knowledge base through conversations

In one sentence: Making content collection and analysis simple, automated, and intelligent.

PlaceholderProduct overview - Data Sources → Knowledge Base → Reports/Ask workflow

Core Value

🔗 Multi-Source Aggregation

• Support 8+ data source types (search, RSS, email, web, tender, etc.)
• Smart URL deduplication to avoid duplicate content
• Automatic version management, always keep the latest valid version

💬 Ask

• Knowledge base AI Q&A (Library Mode)
• Streaming responses with real-time feedback
• Support general conversation mode (General Mode)

🤖 Smart Reports

• Multi-step LLM processing (summarize → analyze → format)
• Each step can use different LLM models
• Scheduled auto-generation or manual trigger

💳 Atomic Billing

• Transaction safe, billing + execution = atomic operation
• Auto-refund on task failure
• Complete transaction audit logs

PlaceholderCore value diagram - 4 module icons

Use Cases

Case 1: Financial News Automation

Requirement: Daily aggregation of financial news from multiple sources, generate industry reports

Features Used:

• Data Sources: Google News + RSS (WSJ, Bloomberg)
• Report Template: 3 steps (summarize → analyze trends → format)
• Schedule: Auto-generate daily at 8:00 AM

Value: Save 90% manual time, from 2 hours to 10 minutes

Case 2: Government Tender Monitoring

Requirement: Monitor IT consulting-related government procurement information

Features Used:

• Data Source: Tender announcement source (CCGP + regional platforms)
• Keywords: "IT consulting", "software development"
• Dedup Strategy: KEEP_OLD (avoid duplicate crawling)

Value: Real-time opportunities, never miss a chance

Case 3: Job Smart Matching

Requirement: Monitor qualifying job positions, auto-match resumes

Features Used:

• Data Source: Search source (Boss Zhipin, LinkedIn)
• Ask: Quick query "What jobs match my skills"
• Report: Weekly recommendation list generation

Value: Improve job search efficiency, precise matching

Core Concepts

1️⃣ Data Source (Source)

Definition: Where to collect content from

Types: Search, RSS, Email, Web scraping, Tender, Google News

Key Settings:

• Schedule strategy (hourly/daily/weekly)
• Dedup strategy (UPDATE/KEEP_OLD)
• Content cleaning (whether to use LLM)

Example: Create RSS data source to monitor "36Kr Tech News", auto-collect every 6 hours

2️⃣ Knowledge Base (Library)

Definition: Categorized storage container for content

Purpose:

• Categorize content by topic
• Used for report generation
• Used for Ask Q&A

Relationship: 1 library can associate multiple sources, 1 source can associate multiple libraries

Example: Create "AI Industry News" library, associate 3 RSS sources + 2 search sources

3️⃣ Report

Definition: Auto-generated analysis reports

Generation Methods:

• Based on report templates
• Multi-step LLM processing
• Scheduled auto-generation or manual trigger

Multi-step Processing:

Step 1: Data aggregation (extract from library)
  ↓
Step 2: Analyze insights (LLM generation)
  ↓
Step 3: Format output (Markdown)

Example: Every Monday at 9 AM, auto-generate "Weekly AI Industry Report"

4️⃣ Credits

Definition: Fee unit for system usage

Acquisition:

• New users get 10,000 credits reward
• Redeem codes for top-up

Consumption Items:

• Data collection: 1-10 credits/time
• Content cleaning: 10-20 credits/time
• Report generation: 50-500 credits/time (depends on model)
• Ask conversation: 1-50 credits/time

PlaceholderCore concepts diagram - Data Sources → Library → Reports/Ask, Credits on the side

Quick Comparison

Next Steps

Ready to get started?

→ Quick Start - Create your first data source in 5 minutes

→ Source Management - Learn about all data source types