Devabase v0.1.0 - Initial Release

The Open-Source Backend for AI Applications

We're excited to announce the first release of Devabase - a lightweight,
self-hostable backend that gives you everything you need to build RAG and
LLM-powered applications.

---

Highlights

Vector Database

• Collections - Create and manage vector collections with customizable
  dimensions and distance metrics (cosine, euclidean, dot product)
• Document Processing - Upload PDF, TXT, MD, DOCX files with automatic
  chunking and embedding
• Semantic Search - Fast similarity search with metadata filtering
• Multi-Collection Search - Query across multiple collections in a single
  request

One-Click RAG

• Instant Chat API - Enable RAG on any collection and get a
  /rag/:collection/chat endpoint
• Configurable LLM - Use OpenAI, Anthropic, or any OpenAI-compatible
  provider
• Source Citations - Responses include relevant document sources with
  relevance scores
• Conversation Memory - Multi-turn conversations with context preservation

Relational Tables

• Dynamic REST API - Create tables and get instant CRUD endpoints
• Foreign Keys - Full referential integrity with ON DELETE actions
• Flexible Queries - Filter, sort, paginate with query parameters
• Import/Export - CSV and JSON support for data migration

SQL Editor

• In-Browser Editor - Monaco-powered SQL editor with syntax highlighting
• Schema Autocomplete - Intelligent suggestions for tables and columns
• Query History - Access and re-run previous queries
• Safe Execution - SELECT-only for security

Developer Experience

• API Documentation - Interactive docs with curl examples
• API Keys - Scoped keys with expiration support
• Webhooks - Real-time event notifications with HMAC signatures
• Multi-Project - Isolated environments for dev/staging/prod

Team Collaboration

• Project Members - Invite team members with role-based access
• Role Permissions - Owner, Admin, Member, Viewer roles
• Email Invitations - Secure invitation flow