I built smart‑context‑memory, a Python library that automatically keeps LLM prompts within token limits by summarizing old messages and retrieving only the most relevant ones.
The repository lives on my GitHub under the handle @sarthaksingh17.
I started the project in April 2026 and last updated it in April 2026.
- Language: Python 3.9+
- Token counting: tiktoken
- Relevance scoring & summarisation: scikit‑learn, numpy