Technology
Indic-Native Astrology LLM
First-party language intelligence built around source-governed classical material. Interpretation coverage spans 30 languages: English, 15 Indic and 14 international.
Where astronomy becomes intelligence
Vedika is a deep tech product at the intersection of artificial intelligence, astronomical computation, and Indic knowledge systems. We build domain-specific AI for computation-heavy traditional knowledge workflows, starting with Vedic astrology.
Our flagship product, Vedika, is first-party Indic-native astrology intelligence spanning Vedic sidereal, Western tropical and KP Krishnamurti contexts plus adjacent documented domains. The published OpenAPI contract lists 559 endpoints across 26 tagged categories. Vedika Standard handles the normal interpretation path; Vedika Swift is the latency-optimized path where included. Response time varies with endpoint, request depth and downstream conditions.
Vedika computes planetary positions, houses and deterministic chart facts before the interpretation layer runs. Generated interpretation remains probabilistic; it is constrained by computed astronomical data and source-grounded rules rather than being presented as mathematical fact.
Founder
Abhishek Raj
CEO & CTO — Founder
Deep tech polymath who built Vedika's Indic-native astrology intelligence, astronomical computation pipeline, and precision layer from scratch. His combination of Investment Banking, Operations, Marketing, and Engineering backgrounds enables him to think across disciplines — from astronomical calculations to business model design.
Previously: Raymond (operations), Pune Angels Network (venture investing). MBA from Symbiosis International University. CFA certification.
Now building domain-specific AI infrastructure for Indic knowledge systems — combining classical Sanskrit texts (BPHS, Phaladeepika, Saravali) with modern deep learning and astronomical computation.
Technology
What the product is built out of, and what each part is and is not claiming.
Technology
Computation Before Interpretation
Vedika separates deterministic chart computation from generated interpretation. Astronomical inputs are computed first; domain rules and source grounding then constrain the interpretation layer.
Technology
XALEN Ephemeris
Current source documents an Apache-2.0 distribution and a JPL DE440 benchmark with results reported by celestial body. Verify the published artifact, version, methodology, scope, results and exceptions before relying on either claim. The dated source also documents divisional-chart, dasha, yoga and ayanamsa fields; customer-callable availability requires current-contract verification.
Technology
Indic NLP
Sanskrit-aware Indic NLP is a product and roadmap focus. Interpretation coverage spans 30 languages, 15 of them Indic. PDF report language coverage is separate and varies by report family.
What's Next
Directions under evaluation. None of the three is a commitment or a statement of current availability.
Roadmap
Quantized Open-Weight Models
The roadmap explores quantized open-weight Indic astrology models and on-device or edge inference with astronomical grounding. This is a research direction, not a commitment or statement that offline capability is currently available.
Roadmap
Indic Language Foundation
The roadmap explores Sanskrit-aware tokenization, script-conscious processing and wider classical-text interpretation coverage. Current language availability must be verified per live surface and contract.
Roadmap
Open Research
The roadmap includes evaluating publishable domain-specific research and potential open-source Indic AI contributions. Publication and release scope remain future work, not guarantees.
By the Numbers
559Published Endpoints
JPL DE440Source-Documented Benchmark*
26Tagged API Categories
2025Founded
30Interpretation Languages
3Astrology Systems
119Live Report Types, of 480 Planned
$12Starting Price/mo
*The JPL DE440 benchmark is described in source documentation. Verify its published artifact, methodology, scope, results and exceptions before relying on it. Endpoint, category and language counts are read from the published OpenAPI contract and the shipped language catalog.
Contact
Where to reach the company, and the pages most people want next.
Vedika Intelligence
Pune, Maharashtra, India
General: [email protected]
Enterprise: [email protected]
Quick Links
Build with Vedika
Review the current public contract, choose a paid plan and request production access.