darkglass 0.4.0

Structured AI system with emebeddable chatbot

Darkglass

Darkglass is a lightweight, embeddable chatbot designed for small colleges, providing prospective students with automated responses while logging interactions for administrative review through a minimal, Python-based stack.

Design

The system is built on a minimalist "bare-metal" architecture, utilizing FastAPI and SQLite to ensure zero-configuration deployment on lightweight Linux VMs without the overhead of external databases. Configuration happens via an optional darkglass.toml file placed in the working directory; supported TOML keys are gemini_api_key and prompt. For example:

gemini_api_key = "mysecretkey"
prompt = "You are a helpful agent for Acme College."

The TOML parser uses Python’s built‑in tomllib, so Python 3.11 or newer is required. To guarantee long-term stability and eliminate dependency churn, the application avoids high-level LLM frameworks in favor of direct REST calls via Python's standard urllib library. The default integration targets Google's Gemini API using the configured API key. Institutional knowledge is managed through a system prompt, either embedded or supplied via TOML, removing the complexity of vector databases and custom search tools. This approach culminates in a cohesive application that serves a lightweight JavaScript widget to prospective students and a secure Google Auth-protected interface for administrators, maintaining a strict, minimal dependency footprint.