MCP 服务器

Vedaksha

Vedaksha MCP Server Astronomical ephemeris and Vedic astrology computation for AI agents via the Model Context Protocol.

README

Vedākṣha — Vision from Vedas

Astronomical ephemeris and Vedic astrology platform. Clean-room Rust implementation with sub-arcsecond planetary precision.

Celestial computation. Agentic precision.

Website · Docs · API Reference · Blog

Quick Start

use vedaksha::prelude::*;

let jd = calendar_to_jd(2024, 3, 20, 12.0);
let chart = compute_chart(
    jd, 28.6139, 77.2090,
    &ChartConfig::vedic()
);

cargo add vedaksha-astro vedaksha-ephem-core

pip install vedaksha

Workspace

Crate	Description
vedaksha-math	Chebyshev polynomials, angle arithmetic, interpolation, rotation matrices
vedaksha-ephem-core	JPL DE440 SPK reader, AnalyticalProvider (VSOP87A + ELP/MPP02), coordinate pipeline, precession, nutation, Delta T
vedaksha-astro	10 house systems, 44 ayanamsha systems (IAU 2006 P03 5th-order precession), aspects, dignities, transits
vedaksha-vedic	27 nakshatras, 3 dasha systems, 16 vargas, 50 yogas, Shadbala
vedaksha-graph	Property graph ontology — 10 node types, 13 edge types
vedaksha-emit	Cypher, SurrealQL, JSON-LD, JSON, embedding text emitters
vedaksha-mcp	Model Context Protocol server — 7 fully functional JSON-RPC tools for AI agents
vedaksha-locale	7-language localization (English, Hindi, Sanskrit, Tamil, Telugu, Kannada, Bengali)
vedaksha-wasm	WebAssembly bindings — 972 KB binary, full chart computation in browser

Plus Python bindings via PyO3 (pip install vedaksha).

Two Ephemeris Providers

Provider	Accuracy	Data	Use Case
SpkReader	Sub-arcsecond	DE440s (31 MB on disk)	Servers, containers
AnalyticalProvider	<15" planets, <1" Moon	Zero files (compiled constants)	WASM, Cloudflare Workers, edge, `no_std`

The AnalyticalProvider uses VSOP87A (Bretagnon & Francou 1988) for planets and ELP/MPP02 (Chapront 2002) for the Moon. All coefficients are compile-time constants — no runtime data files needed.

Computation Pipeline

JPL DE440 SPK → Chebyshev evaluation → ICRS barycentric
  → light-time correction → precession (IAU 2006 P03, 5th-order)
  → nutation (IAU 2000B) → frame bias (ICRS→J2000)
  → aberration → ecliptic coordinates

Or for zero-data environments:

VSOP87A/ELP coefficients (compiled) → Poisson series evaluation
  → heliocentric ecliptic → equatorial rotation → barycentric ICRS
  → same downstream pipeline as above

Delta T: IERS measured table (1620-2025) + Espenak-Meeus predictions to 2050.

Vedic Astrology

First-class Jyotish support — not a Western afterthought.

Nakshatras: 27 lunar mansions with padas, lords, symbols, deities
Dashas: Vimshottari (120-year), Yogini (36-year), Chara (sign-based), Ashtottari (108-year), Narayana (Jaimini)
Vargas: All 16 divisional charts (Rashi through Shashtiamsha)
Yogas: 50 classical combinations (Pancha Mahapurusha, Dhana, Raja, Daridra, etc.)
Shadbala: Complete 6-component planetary strength
Ayanamsha: 44 sidereal systems (Lahiri, Raman, KP, Fagan-Bradley, and 40 more)
Lunar nodes: Mean, True (Meeus 5-term, ~0.09°), and Osculating (<0.03° vs JPL DE441) — KP sub-lord ready
Panchanga: Complete 5-limb day — Tithi (with paksha, lord), Vara (with Rahu/Gulika Kalam), Nakshatra (with deity, yoni, nadi), Yoga (27 astronomical), Karana (60 half-tithis)
Drishti: Graded aspect strengths — Full, ThreeQuarter (75%), Half (50%), Quarter (25%) per BPHS Ch. 26

AI-First Architecture

Every chart computation produces a property graph — not flat structs. AI agents query chart data with Cypher, SurrealQL, or JSON-LD. The MCP server exposes 7 fully functional tools:

compute_natal_chart — Full natal chart with houses, planets, aspects, dignities
compute_dasha — Vimshottari dasha periods to any depth
compute_vargas — Divisional chart positions
compute_transit — Transit positions against natal with aspects
search_transits — Find exact transit events in a date range
search_muhurta — Find auspicious times with quality scoring
emit_graph — Emit chart as Cypher, SurrealQL, JSON-LD, or embedding text

Run the MCP server:

cargo install vedaksha-mcp          # install
vedaksha-mcp                        # stdio (Claude Desktop, VS Code, Cursor)
vedaksha-mcp --http                 # HTTP on port 3100 (Smithery, remote)
vedaksha-mcp --http --port 8080     # custom port
docker run -p 3100:3100 ghcr.io/arthiqlabs/vedaksha-mcp  # Docker

Accuracy

Validated against independent reference ephemerides across 24,000+ oracle data points:

Metric	SpkReader (DE440s)	AnalyticalProvider
Planetary longitude	Sub-arcsecond (avg 1.7")	<15" (avg 3.8")
Moon longitude	Sub-arcsecond	<1" (0.36")
House cusps (10 systems)	Sub-0.001°	Sub-0.01°
Ayanamsha (44 systems)	avg 0.005°	Same (pure math)
Dasha periods	Sum to 120 years ± 0.01 days	Same
Nakshatra boundaries	Reference-accurate	Matches SpkReader at all tested boundaries

Bindings

Platform	Install	Chart Computation
Rust	`cargo add vedaksha-astro vedaksha-ephem-core`	Full pipeline
Python	`pip install vedaksha`	`vedaksha.compute_natal_chart(...)`
WASM	`wasm-pack build crates/vedaksha-wasm`	972 KB, zero data files
MCP	stdio + HTTP transport	7 tools, JSON-RPC 2.0
Docker	`docker run -p 3100:3100 ghcr.io/arthiqlabs/vedaksha-mcp`	HTTP on port 3100

Published Packages

crates.io: 9 crates at v1.5.0
PyPI: vedaksha v1.5.0 (source + macOS arm64 wheel)

License

Business Source License 1.1 (BSL).

Non-commercial use: Free. Personal projects, research, education, internal tools.
Commercial use: $500 one-time per organization. Purchase license.
Converts to Apache 2.0 five years after each version's release date.

See LICENSE for full terms.