Orchid - Research Intelligence Platform — screenshot 1
Orchid - Research Intelligence Platform — screenshot 2
Orchid - Research Intelligence Platform — screenshot 3
Orchid - Research Intelligence Platform — screenshot 4

Orchid - Research Intelligence Platform

ORCHIDE is an agentic, multilingual research intelligence platform for market discovery, financial analysis, and voice-native investigation.

Codex Community Hackathon - Pune

Additional info

How was your experience building with Codex?

It was Pretty good exploring Codex and building a amazing product, which we are planning to ship into future.

Describe your experience using Loops House as the hackathon platform. What worked well, what challenges (if any) did you face, and what improvements would you like to see?

Using Loops House as the hackathon platform was a highly streamlined experience. The interface is intuitive and developer-friendly, which made team coordination and the final project submission process completely frictionless a massive plus when racing against the clock.

Tell us about your overall experience at Codex Community Hackathon Pune.

The overall experience at the Codex Community Hackathon Pune was fantastic! The energy on the floor was electric from start to finish. I deeply appreciated the well-organized structure. he organizers did an exceptional job of maintaining momentum and ensuring teams had everything they needed to succeed.

What could Codex Community improve to create a better experience for participants?

To create an even more impactful experience, the Codex Community could consider integrating brief 'lightning talks' or micro-workshops during the hacking hours. Quick, 15-minute sessions on specific technical bottlenecks (like rapid deployment strategies or API integrations) would give teams an extra edge. Additionally, creating a structured post-hackathon pipeline such as a dedicated alumni network or basic incubation guidance for standout projects would be incredibly valuable for teams looking to scale the solutions they just built.

Team

3 members
  • PR

    prathamesh kolhe

    Agentic Ai Researcher
  • SA

    Sarthak Patil

    Owner
  • UT

    Utkarsh Vidwat

    Data Engineer

Overview

The Core Vision ORCHIDE AI is an institutional-grade, agentic research platform designed to transform raw market data into structured, actionable intelligence. Built to solve the modern crisis of information overload, it provides founders, analysts, and investors with a calm, source-grounded workspace for deep financial discovery. By delegating the heavy lifting of data retrieval, orchestration, and synthesis to an autonomous system, users can focus entirely on high-level judgment and decision-making. The platform operates as a seamless conversational interface, delivering highly accurate, institutional-quality insights without the cognitive clutter of traditional dashboards.

Dual-LLM Orchestration At the heart of the system is a novel Dual-LLM architecture that strictly separates cognitive labor to maximize efficiency, reduce costs, and prevent errors. The "Planner" (powered by Gemini 1.5 Flash) operates entirely in the background, handling complex query decomposition, strategic tool selection, and logical reasoning. Meanwhile, the "Executor" (powered by Sarvam AI) is strictly responsible for high-fidelity user communication, neural voice synthesis, and language translation. This hard architectural boundary ensures that the reasoning engine never interacts directly with the user, and the conversational model never hallucinates unverified research.

Real-Time RAG & Explainable Intelligence To guarantee absolute accuracy, Orchid utilizes a strict, real-time Retrieval-Augmented Generation (RAG) pipeline that explicitly bans unpredictable web scraping. All market intelligence is dynamically pulled from approved, structured APIs—such as GNews, Yahoo Finance, and Moneycontrol—before any LLM processing occurs. Furthermore, every insight delivered by the system is fully explainable, providing the user with explicit source citations, agent reasoning traces, and tool usage logs. To maintain a beautifully minimalist user interface, deep signal metrics and execution logs are routed directly to the developer console.

Multilingual & Voice-Native Discovery Orchid democratizes complex financial intelligence by breaking down language barriers with native support for English, Hindi, Marathi, Tamil, Telugu, Bengali, and Gujarati. Beyond standard text chat, the platform features a dedicated, decoupled voice application powered by the sophisticated Sarvam 'Shubh' neural model. This standalone voice route offers a highly immersive, low-latency research experience, complete with a responsive 3D interactive orb and live-streaming transcripts. This capability allows users to conduct complex, multi-step market investigations conversationally, naturally, and entirely hands-free.

Technical Scalability & Impact Ultimately, ORCHIDE AI redefines how professionals interact with global market data by merging agentic intelligence with an exceptionally crafted, multimodal design. The system is engineered for rapid deployment and enterprise scalability, utilizing a highly performant stack of React, Vite, FastAPI, and robust Redis caching to ensure minimal latency. By prioritizing structured APIs, cost-effective LLM routing, and an uncompromised user experience, Orchid stands out as a production-ready blueprint for the next generation of reliable AI search.