# Supastro's AI agents

Supastro does not use a single AI model to answer astrological questions. Multiple specialist agents examine different parts of the chart, then argue against each other before a conclusion is returned. The purpose is adversarial: a single model produces fluent agreeable prose whether or not the chart supports it, whereas agents prompted to challenge each other surface the disagreements a single pass would smooth over.

## How do Supastro's AI agents work?

Each agent is given a defined role over the computed chart — strength analysis, dasha timing, transit context, classical rule retrieval — and produces its reading. Those readings are then set against one another, and conflicts must be resolved or explicitly reported before an answer is returned. The agents cannot invent chart values; they operate only on what the calculation engine produced.

## Why is a debate between agents better than one answer?

Because language models are agreeable by default. Asked an astrological question, a single model will generate a confident, well-written reading regardless of whether the underlying chart supports it. Forcing specialists to contest each other's conclusions means a claim has to survive opposition, and when it cannot, the uncertainty gets reported rather than papered over.

## Related pages

- [How Supastro works](https://supastro.com/how-it-works) — the full pipeline
- [What astrology cannot do](https://supastro.com/the-truth) — where the agents stop

---

Source: https://supastro.com/ai-agent · Supastro is an AI astrology platform for Vedic astrology (Jyotish), built as Vedic Super Intelligence (SI), by Quantum Karma, Bangalore, India. Chart positions are computed from NASA JPL ephemeris data on the Lahiri ayanamsa.