PORTFOLIO DOSSIER / SOVEREIGN INTELLIGENCE

LOCAL AI · AGENT SYSTEMS · MARKET AUTOMATION · ASTRONOMICAL COMPUTATION

Intelligence that can
reason, act and remain accountable.

Four distinct programmes move from creator-agent orchestration to multilingual consumer assistance, risk-controlled market automation and space-integrated astronomical computation. Each is presented at the maturity supported by its code and blueprint.

PRODUCT LINES04
IMPLEMENTATION EVIDENCECODE + BLUEPRINTS
DEPLOYMENT PRINCIPLEHUMAN-CONTROLLED
STATUSPROTOTYPE PORTFOLIO
AI / 01

Katral AI

A sovereign creator and computational-intelligence programme with two implemented tracks: an autonomous creator workspace and a stock-intelligence prototype. The source portfolio includes a canvas-based agent environment, React/FastAPI applications, authentication, portfolio services, LSTM experimentation and news-sentiment modules.

ENGINEERING PROTOTYPE
01 / CREATOR SYSTEM

Agent factory

Blueprinted role-based clusters for scripting, voice, translation, metadata and multi-channel publishing. Large swarm counts remain a scale target, not a demonstrated capacity.

02 / MARKET TRACK

Stock intelligence

FastAPI and React prototype for market data, trend estimation, portfolio functions, LSTM experiments and real-time news sentiment.

03 / SOVEREIGNTY

Local-first control

Architecture prioritises controllable model routing, private workspaces and reduced dependence on a single external AI provider.

04 / NEXT PROOF

Measured agent run

Package one repeatable creator mission, publish an auditable trace and benchmark quality, latency, cost and human intervention.

MATURITYCODED PROTOTYPES
AVAILABLE EVIDENCECreator UI, React/FastAPI stock app, JWT, portfolio, sentiment and LSTM files
REQUIRED BEFORE LAUNCHUnified product boundary, security review, model evaluation and production observability
AI / 02

Katral Trade Bot

A paper-first, multi-market algorithmic trading workstation built around an authenticated API, browser control plane, pluggable execution, persistent ledger and explicit risk interruption. The updated build separates strategy research from order routing and makes simulated validation the required first operating mode.

FUNCTIONAL LOCAL BUILD
CONNECTIVITY

Three execution paths

Paper, digital-asset venue and desktop-broker adapters share a common interface; no public live endpoint or counterparty relationship is represented.

RISK + RECOVERY

Capital guardrails

Position sizing, protective exits, trailing logic, daily drawdown interruption and broker-ledger reconciliation are implemented.

VALIDATION

Backtest + safety suites

Cost-aware replay and focused tests cover authentication, data state, paper isolation, strategies, risk and execution failure paths.

MATURITYFUNCTIONAL ENGINEERING BUILD
CONTROL PLANEAPI, local ledger, live telemetry, browser console, paper engine and connector framework
REQUIRED BEFORE LIVE CAPITALIndependent strategy, cybersecurity, legal, reconciliation and operations validation
AI / 03

Ramu AI Assistant

A WhatsApp-first personal assistant designed for Indian households and small businesses. Ramu is specified for natural voice and text across English, Hindi and major regional languages, with practical tasks spanning planning, shopping, food, travel, family coordination and business productivity.

BACKEND + LAUNCH BLUEPRINT
01WhatsApp

Meta Business Cloud API as the low-friction user surface.

02FastAPI

Asynchronous webhook and action-orchestration backend.

03Language model

Tool-aware multilingual reasoning with same-language response.

04Action connectors

Planned commerce, food, travel, calendar and local-service actions.

05Payments + data

Razorpay webhooks, SQLModel and PostgreSQL/SQLite fallback.

MATURITYBACKEND PROTOTYPE
LANGUAGE SCOPE14 Indian languages plus Hinglish and mixed-language conversations in the product specification
REQUIRED BEFORE PUBLIC LAUNCHConnector permissions, DPDP controls, payment certification, escalation and safety evaluation
AI / 04

Cosmic Pulse

A research-grade web simulator joining a self-contained astronomical calculation engine with a twelve-node Proof of Pulse orbital-consensus simulation. It resolves geocentric planetary positions from Julian time, applies Lahiri ayanamsha and maps the results into classical Vedic calculation systems.

RESEARCH SIMULATOR
ASTRONOMY

Keplerian resolver

J2000 orbital elements, iterative Kepler equation solving, heliocentric-to-geocentric conversion and sidereal correction.

VEDIC CALCULATION

Four analytical families

Parashara Navamsha and Vimshottari, Jaimini Chara Karakas, Bhrigu age cycles and classical yoga tests.

SPACE SIMULATION

Twelve orbital nodes

Three-second simulated block production, relative node motion, ionisation attenuation, clock-drift modelling and supply state.

DEPLOYMENT

Self-contained stack

Python/Flask API, thread-safe simulator state and client-side orbital calculations reduce dependence on external ephemeris APIs.

SOURCE BASIS / SK\Astro App, SK\Katral AI, SK\trade and SK\Ramu AI Agent. Scientific, financial and performance statements are presented as design intent or simulator behaviour until independently reproduced.
PORTFOLIO NAVIGATION

Continue into commerce, biotech or trusted machine infrastructure.

FLAGSHIP PLATFORM

See how sovereign AI products can inherit identity, settlement, provenance and orbital resilience from the Proof of Pulse stack.

View integrated platform →