A source-available voice-first second brain — an AI memory companion you can call. Think aloud, remember everything, and achieve clarity through conversation. Free for personal and internal business use.
- 🔒 100% Local-First - llama3.2 + local embeddings (Xenova), zero external API calls by default
- 🧠 Voice-First Second Brain - Call and speak your thoughts, AI companion remembers everything
- 📝 Obsidian-Style Notes - Wiki links
[[like this]], knowledge graphs, and markdown editing - 🔗 Semantic Memory Graph - pgvector search retrieves by meaning, not keywords
- 🏠 Self-Hosted First - Full data sovereignty, no subscription required
- 🔌 Pluggable Architecture - Swap LLM/embedding/call providers via config
- 🐳 Docker Ready - One-command deployment with Ollama baked in
- 🪞 Reflective Intelligence - AI surfaces forgotten goals and patterns
Before you begin, ensure you have Docker and Docker Compose installed on your system:
- Docker: Install Docker
- Docker Compose: Install Docker Compose
# Clone repository
git clone https://github.com/abhishekslab/relevel.me.git
cd relevel.me
# Set up environment variables
cp .env.example .env
# Edit .env with your Supabase and CallKaro credentials
# Start all services (web, worker, redis, ollama) with hot reloading
docker compose -f docker-compose.dev.yml -f docker-compose.ollama.yml up --build
# Access at http://localhost:3000Want to run relevel.me on your own infrastructure? Check out our comprehensive Self-Hosting Guide for complete setup instructions, including:
- Docker production deployment
- Supabase project setup and database migrations
- CallKaro voice call configuration
- Custom domain and SSL setup
- Troubleshooting and maintenance
- Local LLM: llama3.2 runs on your infrastructure via Ollama (baked into Docker image)
- Local Embeddings: Xenova/all-MiniLM-L6-v2 runs in-process - zero external API calls
- Privacy by Default: All memory processing happens locally unless you opt into cloud providers
- Offline Capable: Works completely offline after initial model downloads
- Call anytime to speak your thoughts freely
- AI-powered voice conversations with automatic transcription
- Daily reflection calls scheduled by background worker
- Pluggable call providers (CallKaro, Vapi) - swap via config
- Your ideas become connected nodes in a knowledge graph
- pgvector semantic search retrieves memories by meaning, not keywords
- AI surfaces forgotten goals, patterns, and contradictions
- Context-aware recall through natural conversation
- Full data sovereignty - your thoughts stay on your infrastructure
- No subscription required for self-hosted deployment
- Pluggable architecture: swap LLM providers (local/OpenRouter), embedding models, call services
Hosted Version: Coming Soon
Currently, relevel.me is available as a self-hosted solution. A fully managed hosted version is in the works.
Self-Hosted (Available Now)
- Full control over your data and infrastructure
- No subscription required
- Estimated cost: $5-30/month (for cloud hosting)
- Setup time: 30-60 minutes
- Full code access for customization
For detailed self-hosting instructions, see docs/SELF_HOSTING.md
- Architecture Guide - Tech stack, monorepo structure, and system architecture
- Notes System - Obsidian-style notes with wiki links, knowledge graphs, and voice integration
- Self-Hosting Guide - Complete guide to running relevel.me on your own infrastructure
- Onboarding Flow - User journey audit and flow analysis
- Gaps and Fixes - Implementation notes and fixes applied
Copy .env.example to .env and configure:
We welcome contributions from the community! 🎉
- Read CONTRIBUTING.md for guidelines
- Report bugs by opening an issue
- Suggest features in GitHub Discussions
- Submit PRs for bug fixes or new features
- Add call providers - See docs/PROVIDERS.md
- Improve docs - Documentation PRs are highly valued
Please read our Code of Conduct before contributing.
Found a security vulnerability? Please report it responsibly:
- DO NOT open a public issue
- Read our Security Policy for reporting instructions
- Use GitHub Security Advisories for private reporting
Elastic License 2.0 - see LICENSE for details.
TL;DR: Free for personal and internal business use. Cannot offer as a hosted service to third parties. Derivative works must be shared under the same license.
Copyright (c) 2025 relevel.me contributors
For commercial licensing inquiries: hello@abhishekslab.xyz
- 📖 Documentation: docs/
- 🐛 Issues: GitHub Issues - Report bugs
- 💬 Discussions: GitHub Discussions - Ask questions, share ideas
- 🔐 Security: Security Policy - Report vulnerabilities
- 🏠 Self-Hosting: Self-Hosting Guide - Run your own instance
- Setup Guide - Development setup
- Provider Guide - Call provider configuration
- Self-Hosting Guide - Production deployment
- Architecture Docs - Codebase overview