Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs-chat.die-coaching-akademie.de/llms.txt

Use this file to discover all available pages before exploring further.

​
Module reference

PathResponsibility
app.pyFastAPI app, public routes, admin routes, static files, analytics hook
models.pyPydantic models for chat, history, state, and health responses
core/config.pyPaths, environment helpers, system prompt, constants
services/chat_service.pyChat answer logic, knowledge-base search, deterministic rules, OpenAI calls
services/analytics_service.pyJSONL analytics, scrubbing, summaries
sync_website_content.pyWebsite content sync into data/website_knowledge_base.json
static/chatbot.jsEmbeddable widget behavior
static/chatbot.cssWidget styling
templates/admin_dashboard.htmlProtected admin dashboard page

​
Change guidance

Prefer services/chat_service.py. Keep app.py as the HTTP entrypoint and add focused regression tests in tests/test_app.py.
Prefer data/knowledge_base.json for curated content. Use data/CONTENT_GUIDE.md for field expectations.
Use static/chatbot.js for behavior and static/chatbot.css for styling. The customer site only embeds the script.
Use services/analytics_service.py for stored event shape and summary functions. Keep privacy constraints intact.

​
Example: deterministic answer order

answer_chat currently tries deterministic handlers before OpenAI: 
for response in (
    build_payment_terms_response(payload.message),
    build_price_response(payload.message),
    build_infoabend_response(payload.message),
):
    if response:
        store_cached_response(payload.message, response, payload.history)
        return response
This order matters because known high-risk topics like payment and price should not be improvised by the model.