How I Generate Answers

From chunks to a cited, readable answer

Once the vector search returns the most relevant chunks from the rulebook, I need to turn them into a coherent answer. That is the synthesis step, and here is how it works:

Loading diagram...

Each step has a specific job. The prompt constructor injects the chunks in a structured way, tells the model which citation markers to use, and sets the output language.

Citation system

Every answer I produce links back to a source. The markers work as follows:

These markers become clickable links in the web interface. [PDF1, p.34] opens the PDF at page 34. [T1] opens the BGG thread.

Prompt templates

I manage 6 category-specific synthesis templates for Tier 1:

• synthesis-tier1-yes_no-normal.yml
• synthesis-tier1-rule_explanation-normal.yml
• synthesis-tier1-procedural-normal.yml
• synthesis-tier1-overview-normal.yml
• synthesis-tier1-edge_case-normal.yml
• synthesis-tier1-multi_question-normal.yml

The extraction step classifies the question into one of these 6 categories, and I route to the matching template. Edge-case questions get a different prompt structure than simple yes/no questions.

Why OpenRouter?

I route AI calls through OpenRouter rather than calling OpenAI or Anthropic directly. This gives me:

• Single API key for multiple models
• Automatic fallback between GPT-4o and Claude if one is down
• Cost tracking across both providers

Response normalizer

After synthesis, the raw AI output goes through a normalizer that:

• Strips leftover AI reasoning artifacts
• Ensures citation numbers are consistent with the source list
• Converts PDF URLs to the right format for each channel (Telegram vs Web)