Westace10/techtrvlr-agent

An AI-powered content generation agent for creating tech and travel-focused social content for African audiences using an MCP-inspired architecture. # Techtrvlr Agent An AI content-generation agent, built around an MCP-style agent abstraction, that researches a topic in real time and turns it into ready-to-post social content — text, hashtags, and an image — for tech/travel and cybersecurity-awareness themes aimed at African audiences. ## How it works `main.py` defines an abstract `MCPAgent` interface with a concrete `ContentGeneratorAgent`: it pulls context from a fixed set of African tech news sources (TechCabal, Disrupt Africa, and others), passes that context to OpenAI to draft the post text and hashtags, and generates an accompanying image via DALL-E. `intelligence_engine.py` / `enhanced_intelligence_engine.py` extend this into a research-brief generator that takes a user profile (name, focus areas, tone) and produces an intelligence brief plus platform-specific spin-offs (a LinkedIn post, a Twitter thread) from the same source research. `find_ig_account.py` and `ig_debug_script.py` are small utilities for locating and verifying your own Instagram Business Account ID via the Facebook Graph API, for the posting step. ## Example output A generated cybersecurity-awareness post (image generation and hashtags included, image URL omitted since DALL-E links expire after an hour): > Your work computer is like your kitchen - it needs regular cleaning to stay healthy! Clearing your browsing history is one simple way to keep things tidy and secure. > > 💡 Quick Tip: Take a minute to clear your browsing history today - your computer will thank you! > > `#CyberSecurity #StaySafeOnline #CyberAwareness #TechTips` A generated intelligence brief, from a test user profile focused on fintech, condensing real news items (a PayPal/Paidy acquisition, an EU crypto-transfer regulation proposal, Google Pay's bank partnerships, an a16z crypto fund) into a summary with analysis and forward-looking implications — then repackaged from the same research into a separate LinkedIn post and Twitter thread. ## Getting started Require …