Vedākṣha

The canonical introspection mirror at https://vedaksha.net/api/mcp answers initialize, ping, tools/list, get_version, and tools/call over JSON-RPC 2.0. Tool schemas mirror the live tool_definitions() in the Rust source; a drift-guard test (snapshot_matches_current_tool_definitions) fails CI if the published tools/mcp-tools.json snapshot diverges. To run these tools, self-host the library or build from source — the introspection mirror is read-only by design.

# Discover the canonical 12-tool catalog
curl https://vedaksha.net/api/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

For a fast no-install evaluation, a small technology showcase with 4 agent-friendly demo tools is hosted at mcp.vedaksha.net/mcp (free tier, anon, IP-rate-limited) — separate surface, not the full library.

// vedaksha-mcp · compute_dasha
// Sources: BPHS Ch.46–47, Jaimini Sutras
pub enum DashaSystem {
    Vimshottari, // 120-yr · 9 lords · moon
    Ashtottari,  // 108-yr · 8 lords · moon
    Yogini,      // 36-yr · 8 yoginis · moon
    Chara,       // Jaimini · lagna-based
    Narayana,    // Jaimini · lagna-based
}

// vedaksha-emit · CypherEmitter
// 10 node types · 13 edge types
MERGE (c:Chart {id:"chart_2024_03_20"})
MERGE (sun:Planet {name:"Sun"})
MERGE (sun)-[:Occupies]->(:Sign {name:"Pisces"})
MERGE (sun)-[:PlacedIn]->(:House {number:4})
MERGE (sun)-[:InNakshatra]->(:Nakshatra {name:"U.Bhadrapada"})

// then traverse like any graph
MATCH (m:Planet {name:'Mars'})
   -[:Aspects]->(s:Planet {name:'Saturn'})
RETURN m, s

Deterministic IDs mean re-emitting the same chart is idempotent. Multi-chart queries (matchmaking, transit synastry, RAG over a chart corpus) become graph patterns instead of bespoke loops.

Run cargo test --release from a fresh clone — 870 unit tests plus 8,700 oracle rows against JPL Horizons run in under a minute.

// MCP input schema for compute_natal_chart —
// the entire surface area an agent or client sees.
{
  "required": ["julian_day", "latitude", "longitude"],
  "properties": {
    "julian_day":   { "type": "number" },
    "latitude":     { "type": "number" },
    "longitude":    { "type": "number" },
    "house_system": { "type": "string" },
    "ayanamsha":    { "type": "string" }
  }
}

// fields not in the schema, by construction:
//   name, email, birth_date, address, phone,
//   user_id, ip_address, device_id

// every output graph is tagged at construction:
ChartGraph {
  …,
  classification: DataClassification::Anonymous,
}
// upgrade to Pseudonymized / Identified is an
// explicit caller decision, never the default.

Names of people never enter the engine. Every output graph carries a DataClassification tag so downstream systems can enforce retention and access policy. GDPR and HIPAA-adjacent compliance is a property of your data layer, not the math layer.

Sources cited inline in the source tree: Meeus (Astronomical Algorithms), Bretagnon & Francou (VSOP87A), Chapront (ELP/MPP02), Santhanam (BPHS), JPL DE440 / DE441.