Clariq

Overview

Clariq is your AI-powered sales research partner that delivers clear, actionable insights on prospects, markets, and opportunities—helping you close smarter, faster, and with confidence. It automates data collection, synthesizes findings using state-of-the-art AI, and turns them into concise, decision-ready briefs. Use it from the web app or programmatically via the API to trigger research jobs, track progress, and retrieve structured results.

Key Features

• Research jobs with async orchestration: queue execution, poll status, and fetch results via a clean REST API
• Multi-source data collection with Exa search integration
• AI summarization and report generation powered by Cerebras models and prompt templates
• Voice-enabled workflows using LiveKit for real-time transcription and interactive sessions
• Secure authentication with Appwrite session JWTs
• Modern, scalable stack: FastAPI backend, Next.js frontend, cloud-ready deployment

LIVE LINK 🌱 | YOUTUBE DEMO 🎥

Tech Stack

Backend

• Framework: FastAPI, Python, Uvicorn
• Database: Appwrite (JWT auth, collections)
• AI/LLM: Cerebras Cloud SDK (llama-4-scout-17b-16e-instruct)
• Search: Exa API (neural search, company discovery)
• Voice: LiveKit Agents (Deepgram, Silero VAD, OpenAI TTS)

Frontend

• Framework: Next.js 15, React, TypeScript
• Styling: Tailwind CSS, shadcn/ui
• Auth: Appwrite Client SDK
• Real-time: LiveKit Client SDK

DATAFLOW DIAGRAM

API

Backend API Endpoints

Method	Endpoint	Description	Auth	Response
GET	/health	Liveness/readiness probe	❌	200 OK
GET	/	Root message	❌	200 OK
GET	/v1/auth/test-connection	Public connectivity check	❌	200 OK
GET	/v1/auth/me	Get current authenticated user	✅	200 OK
POST	/v1/research/execute/{job_id}	Queue execution for a research job	✅	202 Accepted
GET	/v1/research/status/{job_id}	Get status for a research job	✅	200 OK
GET	/v1/research/health	Research service health check	❌	200 OK

Note: Base URLs: Local http://localhost:8000 | Production https://api.clariq.app
All endpoints prefixed with /v1 except root and health. Authentication via Appwrite JWT: Authorization: Bearer <token>

Frontend Routes

Route	Description	Auth
/	Landing page with features and CTA	❌
/signin	Authentication page	❌
/dashboard	Main dashboard	✅
/dashboard/research	Research jobs interface	✅
/dashboard/voice	Voice session interface	✅
/test-auth	Authentication testing page	✅

Note: Deployed at https://clariq-agent.vercel.app

DATABASE SCHEMA

Future Scope

AI-Powered Voice Agent (In Development)

Currently, the voice agent feature is under development. We are building an intelligent conversational AI system that will:

• Research-Driven Conversations: Create public voice rooms where AI agents can interact with clients based on completed research data
• Context-Aware Discussions: The agent will leverage research insights (company profiles, market analysis, competitor data) to have informed, personalized sales conversations
• RAG Integration: All voice conversations will be transcribed in real-time and stored in a Retrieval-Augmented Generation (RAG) system
• Continuous Learning: Transcripts feed into the context database, allowing the LLM to learn from previous interactions and improve responses over time
• Session History: Users can review past conversations, extract key talking points, and track engagement metrics

Technical Implementation:

• LiveKit for real-time audio streaming
• Deepgram for speech-to-text transcription
• Vector database for conversation embeddings
• Cerebras LLM for context-aware response generation
• OpenAI TTS for natural voice synthesis

This feature will enable sales teams to have AI agents conduct preliminary discovery calls, answer client questions, and gather insights autonomously while maintaining conversation quality and context throughout multiple sessions.

---

Development: See backend/README.md and client/README.md for setup instructions.

Thank You

---

Built with ❤️ by the Clariq Team