AI should explain the knowledge, not manufacture the chart.

• Explain calculated Jyotish facts in accessible language.
• Answer questions using supplied chart, Life Map, Focus, or Together context.
• Translate or adapt explanations across supported languages and explanation depths.
• Retrieve and synthesize relevant source material where the product provides source-backed reasoning.
• Generate reflective prompts or guidance within configured boundaries.

Kaaladarshi is designed around a layered flow:

1. Birth information is normalized.
2. A deterministic calculation engine produces chart and timing facts.
3. Structured Jyotish rules may derive Focus, period character, and other signals.
4. Source material may be retrieved where relevant and available.
5. AI receives selected context and helps communicate the result.

The AI-generated wording can change without changing the underlying chart. Users should be able to inspect the chart and reasoning where the product provides that capability.

AI responses can be incomplete, overly general, inconsistent, mistranslated, or incorrect. A response can misunderstand a question, fail to apply context correctly, omit an important factor, or produce wording that sounds more certain than the evidence supports.

Different Jyotish schools may interpret the same configuration differently. Birth-time uncertainty, calculation settings, source availability, and selected period context can materially affect an interpretation.

Kaaladarshi may support multiple languages with different levels of translation completeness. English fallback may appear where a localized string or explanation is unavailable.

Jyotish terminology is not always translated literally. Sanskrit-derived or established terms may remain transliterated to preserve meaning. AI-generated translation can make mistakes, so users should compare technical terminology with the underlying chart when precision matters.

Kaaladarshi should avoid fear-based certainty, coercive language, and instructions that put users at risk. It should not tell a user to stop medication, avoid medical care, remain in danger, make an investment, sign a contract, quit a job, end a relationship, or take another consequential action solely because of a Jyotish reading.

For urgent safety, health, legal, or financial concerns, use appropriate professional or emergency resources rather than relying on Kaaladarshi.

Practical, reflective, classical, or traditional guidance may be presented where the service has a reasonable basis to do so. Traditional practices should be distinguished from deterministic calculation and from AI-generated interpretation.

Future provider or service suggestions should not change the underlying Jyotish reasoning merely because a commercial provider is available.

Kaaladarshi may use one or more third-party AI models and may change models, vendors, routing, prompts, retrieval systems, or model settings over time. A particular model is not guaranteed to remain available.

Where AI providers process user questions or contextual information, that processing is governed by Kaaladarshi's Privacy Policy and applicable provider arrangements.

Kaaladarshi uses regression testing and may use curated evaluation examples, user feedback, aggregate product signals, and known chart cases to improve quality. Historical alignment is not treated as proof that future prediction is certain.

Feedback should help identify incorrect calculations, poor reasoning, unclear language, unsafe guidance, and translation problems rather than simply rewarding answers that sound agreeable.

You decide what context to provide, whether to save it, and how much weight to give any interpretation. Treat Kaaladarshi as one input among lived experience, evidence, trusted relationships, professional advice, and your own judgment.

For more on calculation versus interpretation, see Methodology.