AI Landscape Monitor

An automated intelligence briefing engine to stay ahead of the curve.

Open Source (MIT License) Blog Post

Overview

The Challenge: Information Overload

The explosive growth of AI research publications has made manual tracking nearly impossible. According to our arXiv growth tracker, the volume of AI-related papers on arXiv alone has grown exponentially—with thousands of new submissions monthly across computer vision, natural language processing, machine learning theory, and robotics. This torrent of information, while exciting, creates a signal-to-noise problem that only automation can effectively address.

The AI Landscape Monitor is a web-based tool designed to automate the creation of weekly intelligence briefings on the state of Artificial Intelligence. It leverages large language models (LLMs) to gather and summarize recent information across key domains, allows for manual curation and content synthesis, and formats the final output into a clean, readable HTML digest.

The primary goal is to drastically reduce the time and effort required for individuals and teams—such as researchers, developers, and product managers—to stay current with the rapidly evolving AI landscape. By automating the gathering, summarization, and formatting processes, the tool transforms hours of manual research into minutes of configuration.

Open Source Coming Soon! This tool will be released as open source. Read the announcement blog post for more details.

Core Features

Automated Content Generation

The tool automatically generates content for four distinct, pre-defined sections:

  • 📚Research Highlights: Summaries of novel AI research papers with meaningful contributions.
  • 🌐News & Trends: Key industry news, model releases, partnerships, and policy updates.
  • ⚙️Tools & Resources: New open-source projects, datasets, and frameworks.
  • 💡Perspectives & Ethics: Analysis of AI safety, policy, and societal impact discussions.

Selectable AI Engines

Choose between different backend search engines for content generation to compare outputs or leverage specific model strengths:

  • Perplexity (sonar-pro): Optimized for search-augmented, factual responses with citations.
  • Anthropic (claude-sonnet-4-5): Known for speed, intelligence, and strong analytical capabilities.
  • OpenAI (gpt-4.1): Web search enabled models with live internet access and citation support.
  • Brave: Privacy-focused search with freshness filtering, extra snippets, and section-optimized queries.
  • SerpApi: Google Search Results API with knowledge graph, answer box, and section-optimized queries.
  • ⚡Set a global engine or assign different engines per section (mixed mode).

📊 Performance Notes (6-day search tests):

  • Perplexity: Fast execution with diverse source discovery for tools and resources.
  • Claude: Strong general performance; best for research papers with effective fallback to extended time windows.
  • OpenAI (gpt-5): Slower but comprehensive coverage with detailed analysis.
  • Brave: Consistent 10 results with extra snippets; uses section-optimized queries for improved relevance and recency.
  • SerpApi: Largest result counts (2700-415K total); includes Knowledge Graph, Answer Box, and Related Questions with section-optimized queries.

Customizable Content

Tailor the briefing's focus to your specific needs:

  • ✏️Prompt Editing: Modify the default prompts for each section to focus on specific topics, industries, or regions of interest.
  • 🔄Section Toggling: Enable or disable any of the four sections to create a more focused digest.
  • 🌍Cross-Domain Adaptability: Use the DOMAIN_KEYWORD environment variable to adapt search queries to any domain—e-government, development economics, climate tech, cybersecurity, or beyond.

Manual Curation & Synthesis

Enhance your digest with hand-picked resources:

  • 🔗Add Custom Links: Include your own URLs with titles and categories to manually curate important resources.
  • ✨Auto-Synthesis: The tool scrapes URL content and uses your selected AI engine to generate concise, one-paragraph summaries automatically.
  • 📂Organize by Category: Group your manual links under custom categories for better organization.

Configurable Citation Styles

Choose the citation format that suits your presentation needs:

  • 🔗Inline Style: Clickable, bracketed citation numbers (e.g., [1]) are placed directly in the text for quick reference.
  • 📚Section Style: Citation numbers appear in the text, with a full, clickable list of sources provided at the end of each section for deeper review.

Session Management

Your work is automatically saved and preserved:

  • 💾Server-side session saves all customizations (API choice, prompts, manual links).
  • ⏱️Configure your digest now and generate it later without losing your work.
  • 🔄Revisit and regenerate with updated sources or parameters.

Export & API Access

Multiple ways to access and share your briefings:

  • 📄Save as HTML: Download self-contained .html files with all formatting preserved for easy distribution and archiving.
  • 🔌REST API: Programmatic access for scheduled automation and integration into workflows.
  • ⚡Parallel Processing: Toggle between parallel (faster) or sequential (debugging-friendly) processing modes.
  • 🚀Auto-Publishing: Optionally post generated digests to SharePoint Online, WordPress, Confluence, or other platforms via configurable APIs and webhooks.

Workflow

1. Configure Briefing
INPUTS

Select AI Engine

Customize Prompts

Add Manual Links

2. Process & Generate
AI MONITOR ENGINE

Executes Multiple Queries

Gathers & Summarizes Data

Formats Citations

3. Receive & Share
OUTPUT

Curated HTML Digest

Ready for Distribution

Archive-Ready

4. Auto-Publish (Optional)
SHAREPOINT ONLINE

Automatic Posting

Configurable Venues

Scheduled Delivery

User Interface

AI Landscape Monitor User Interface

Clean, intuitive interface for configuring and generating intelligence briefings

API Documentation

Programmatic access to generate intelligence briefings with full customization options.

GET /health

Health check endpoint to verify the service is running.

curl http://localhost:8000/health

Response: {"status": "healthy"}

GET POST /api/session

Manage session configuration (API choice, citation style, enabled sections, etc.).

GET - Retrieve current session:

curl http://localhost:8000/api/session

POST - Update session configuration:

curl -X POST http://localhost:8000/api/session \
 -H "Content-Type: application/json" \
 -d '{"api_choice": "claude", "citation_style": "inline"}'
POST /api/synthesize-url

Generate a summary for a given URL.

curl -X POST http://localhost:8000/api/synthesize-url \
 -H "Content-Type: application/json" \
 -d '{"url": "https://example.com/article", "id": "link-123"}'

Response: {"synthesis": "Article summary text..."}

POST /api/generate-digest

Generate a digest using current session configuration.

curl -X POST http://localhost:8000/api/generate-digest \
 -H "Content-Type: application/json"

Response: {"digest": "HTML content..."}

POST /api/external/generate

External API for generating digests with custom configuration. Requires API key via X-API-Key header.

Basic request:

curl -X POST http://localhost:8000/api/external/generate \
 -H "Content-Type: application/json" \
 -H "X-API-Key: your-api-key" \
 -d '{"api_choice": "perplexity", "citation_style": "inline", "days_delta": 7}'

Full configuration example:

curl -X POST http://localhost:8000/api/external/generate \
 -H "Content-Type: application/json" \
 -H "X-API-Key: your-api-key" \
 -d '{
   "api_choice": "perplexity",
   "citation_style": "inline",
   "days_delta": 14,
   "enabled_sections": {
     "research": true,
     "news": true,
     "resources": false,
     "perspectives": true
   },
   "section_api_choices": {
     "research": "claude",
     "news": "perplexity",
     "resources": "perplexity",
     "perspectives": "claude"
   },
   "use_parallel": true,
   "manual_links": [
     {"url": "https://example.com", "title": "Custom Link", "category": "AI Research"}
   ]
 }'

Custom prompts example:

-d '{
  "queries": {
    "research": "From the past 30 days, select up to 5 AI research papers...",
    "news": "From the past 30 days, what are the top 5 AI developments..."
  }
 }'

Response: Full HTML digest page (Content-Type: text/html)

Configuration Reference

Request Parameters

  • api_choice: perplexity | claude
  • citation_style: inline | section
  • days_delta: number (default: 7)
  • use_parallel: boolean (default: true)
  • section_api_choices: object with per-section engine selection
  • enabled_sections: object with boolean values
  • queries: object with custom prompts
  • manual_links: array of link objects

Link Object Format

{
  "url": "https://...",
  "title": "Link Title",
  "category": "Category",
  "synthesis": "Auto-generated summary"
}

Sample Reports

Explore sample outputs to see the AI Landscape Monitor in action across different domains and use cases.

AI Weekly Digest

A sample weekly intelligence briefing. Engine config: Claude (research, news, resources) + Perplexity (perspectives).

View Sample

AI Weekly Digest v2

Mixed engine configuration demonstrating parallel processing. Engine config: Claude (research) + OpenAI (news, resources) + Perplexity (perspectives).

View Sample

AI Weekly Digest v3

Optimal engine pairing by content type. Engine config: Claude (research, resources) + OpenAI (news, perspectives).

View Sample

eGov Use Case

Demonstrates cross-domain adaptability with e-government, digital public infrastructure, and GovTech applications.

View Use Case

eGov Digest Sample

A sample digest output customized for digital government, featuring e-gov research, policy updates, and GovTech innovations.

View Digest

Note: These samples demonstrate the tool's flexibility. Use the customizable prompts to adapt the briefing for any domain or industry.

Future Directions

Planned enhancements to expand the AI Landscape Monitor's capabilities.

Scheduled Reports

Automate report generation and delivery on custom schedules.

  • • Cron-like scheduling (daily, weekly, custom intervals)
  • • Email delivery with HTML attachments
  • • Webhook notifications for Slack, Teams, etc.

Custom Templates

Save and reuse report configurations.

  • • Save engine, section, and prompt configurations
  • • Template library with one-click application
  • • Share templates with team members

Report Composer

Mix and match sections from multiple reports produced with different search engines.

  • • Select sections from any report (historical or same-day)
  • • Combine multiple sources into a single composite
  • • Ideal for organizational sharing and synthesis

Multi-Engine Synthesis

Evolution

Automatically run a topic through multiple search engines and synthesize into one unified digest—no duplicates.

  • • Single query runs across Perplexity, Claude, and OpenAI
  • • Automatic duplicate detection and removal
  • • Synthesizes unique insights from each engine
  • • Produces one comprehensive, de-duplicated briefing
By Ng Chong | UNU Campus Computing Centre