🪷 विवेक

viveka

Vivekachudamani (Shankaracharya)

Discriminative Wisdom — seeing through maya

Input/output filtering for LLM agents. Content filtering, PII detection, topic classification.

Formerly agent-filter — Part of the Vedic Arsenal: 100 production-grade Python libraries for LLM agents, each named from the Vedas, Puranas, and Mahakavyas.

The ancient Viveka principle from Vivekachudamani (Shankaracharya) finds its modern expression in this library.

Just as the Vedic sages understood that every phenomenon in the universe follows deep patterns — patterns of creation, maintenance, and dissolution — viveka applies this wisdom to LLM agent engineering.

The concept of विवेक (Discriminative Wisdom — seeing through maya) speaks directly to the technical problem this library solves. When the sages codified this principle in Vivekachudamani (Shankaracharya), they were describing not just a spiritual truth but a computational truth that would take humanity millennia to rediscover in silicon.

This is not coincidence. The universe has one nature. The Vedas described it first.

flowchart LR
    A[Input] --> B[viveka]
    B --> C{Process}
    C -- Success --> D[Output]
    C -- Error --> E[Handle]
    E --> B
    style B fill:#6b21a8,color:#fff

pip install viveka

Or from source:

git clone https://github.com/darshjme/viveka.git
cd viveka && pip install -e .

from viveka import *

# See examples/ for full usage

viveka is one of 100 libraries in darshjme/arsenal — each named from sacred Indian literature:

Each library solves one problem. Zero external dependencies. Pure Python 3.8+.

1. Fork the repo
2. Create feature branch (git checkout -b fix/your-fix)
3. Add tests — zero dependencies only
4. Open a PR