Why I Started Using Easy Geography Prompts
I worked on a trivia platform for about four years, building geography questions for mobile apps and classroom tools. The bottleneck was always the same: writing good prompts that didn't repeat patterns or accidentally give away answers. Easy Geography Prompts came up on a forum thread, I downloaded it, and it turned out to be exactly the kind of thing I wished existed when I was building question banks by hand. The tool itself is a prompt generator for geography-based educational content. You feed it parameters like region, difficulty tier, question type, and format style, and it outputs ready-to-use prompts in JSON or CSV. It doesn't guess the answers for you. It structures the question so you can slot it into whatever system you're using. That distinction matters more than people usually realize.
Getting Started with Easy Geography Prompts
Download it from the official repo at github.com/easygeography/prompts. The latest release is version 2.4, released in March 2026. It runs on Python 3.10 or newer. You don't need an API key, which is one of the reasons it stays useful in classrooms where network access is unreliable. After installation, run the setup script. It asks for your output directory and preferred template style. I use the structured-quiz style because it maps cleanly to SQLite imports. Once configured, you can generate a batch with a single command: geoprompts generate --region=SE_Asia --difficulty=medium --count=50 --output=asia_batch.csv
That's the core workflow. Most people stop there and wonder why their question sets look flat. The real work happens in the refinement step, which the documentation barely covers.
Get the Full Details
What Beginners Get Wrong
The biggest issue I see is that people treat the output as final. The prompts it generates are structurally sound but geographically lazy if you don't edit them. Here's a concrete example: I generated a batch of 100 prompts for Sub-Saharan Africa and pulled five at random. Three of them asked about the same country without variation in question type. One used "longest river" as the anchor for Nigeria, which is factually incorrect and would have burned credibility immediately. The workaround is simple but time-consuming. Run a validation pass through a script I wrote that cross-references every country mentioned against the ISO 3166-1 alpha-2 list and flags any capital city mismatches. It catches about 90% of the obvious errors. The remaining 10% require manual checking against a source like the CIA World Factbook. Another counter-intuitive thing: difficulty scaling in Easy Geography Prompts isn't controlled by the prompt complexity itself. It's controlled by the number of distractor options and how specific the geographic detail is. A question about "where is Jakarta located" and a question about "which strait separates Sumatra from Java" might both be tagged "medium," but the second one demands a significantly different knowledge tier. If you're building adaptive quizzes, you need to remap the difficulty labels yourself based on your student population's baseline.
The Edge Case That Almost Drove Me Away
About six months after adopting the tool, I tried generating prompts for micronations and disputed territories. The tool has no built-in handling for places like Kosovo, Taiwan, or Somaliland. It either skips them entirely or defaults to the most recognized sovereign state boundary, which produces politically charged or outright wrong results depending on your audience. I spent two days writing a custom filter that maps these regions to a configurable dispute-awareness flag, so the prompts can be generated with contextual notes attached rather than silently erasing them. There's no update to this in the main branch as of version 2.4. The maintainers are aware of it. If you're working in a context where territorial sensitivity matters, you'll need that filter or an equivalent patch.
Advanced Usage: Chained Prompt Generation
One feature that isn't well documented is the chaining capability. You can feed the output of one generation run into another run as a reference set, and Easy Geography Prompts will avoid repeating the same geographic entities at the same difficulty level. This is useful when you're building a full curriculum rather than a one-off quiz. The command looks like this: geoprompts generate --chain=previous_batch.csv --region=South_America --difficulty=easy --count=30 It reduces duplicate country coverage from about 40% down to under 8% in my testing. The trade-off is that generation time increases by roughly 3x because it has to parse and index the previous batch before generating new prompts. On a standard laptop, a 30-prompt chained run takes about 90 seconds compared to 30 seconds for a standalone run.
Limitations That Actually Matter
Easy Geography Prompts does not handle maritime boundaries, exclusive economic zones, or temporal geography. If you need prompts about historical borders or current territorial disputes beyond the micronation issue, this tool won't help you. The output is static and current as of the last dataset refresh, which for version 2.4 appears to be late 2025. It also doesn't support non-Latin scripts natively in the prompt templates. If you're generating content for Arabic or Japanese-language classrooms, you'll need to post-process the output through a translation layer or modify the template engine directly. I wrote a simple wrapper that routes the CSV through a local LibreOffice conversion step, which adds about 5 minutes to a 100-prompt batch but keeps everything in-house without API dependencies. For most users building standard multiple-choice or short-answer geography quizzes, this is more than adequate. For specialized applications involving cartography, geopolitical analysis, or multilingual deployment, expect to invest additional time in customization or consider pairing it with a dedicated GIS question builder instead.
