Changelog

00 // POLICY

Changelog Policy

This page documents every release of collaborAItr. We follow Semantic Versioning — major.minor.patch — with pre-release labels for beta and release-candidate builds.

v1.0.0-beta.2

BETA

MARCH 19, 2026

A major beta update for collaborAItr focused on public sharing, billing accuracy, admin controls, profile routing, and the new Settle It With AI experience.

ADDED

Launched the public Settle It With AI experience with shareable result pages at /settleit
Added GPT-5.4 model support
Added in-app feedback submission and richer bug-report tools
Added a secure public contact form email endpoint

IMPROVED

Account pages now support direct /account/* routes and use Clerk's hosted profile management for more reliable profile and billing flows
Shared conversations and Settle It links now have improved social preview metadata with app-domain OG routing fixes
Pricing is now available to signed-out visitors, and billing screens now use the effective database-backed plan state
Conversation sharing and routing were improved with canonical /convo/:id URLs and public-by-default shared conversations

FIXED

Hardened Clerk auth and billing synchronization so plan tiers, usage limits, model access, checkout gating, and subscription recovery stay consistent
Fixed OpenRouter token and usage accounting so streamed chats preserve authoritative counts end-to-end
Restored correct Basic-tier access for student users
Fixed pricing and subscription warning banners to reflect the effective database-backed plan

v1.0.0-beta.1

BETA

MARCH 4, 2026

The first public beta of collaborAItr — a multi-AI parallel chat platform that lets you query up to 4 AI models simultaneously and compare their responses in real time.

CORE CHAT

Multi-model parallel chat — query 1–4 models at once with independent streaming responses
Color-coded model tabs with provider logos
Per-model conversation threads with preserved context
Sticky follow-up bar for continuing conversations
Thinking/reasoning mode with collapsible sections for supported models
Response consolidation — merge all model answers into a unified summary
Web search integration with URL crawling and content extraction
File creation via tool calling — models can generate downloadable files

AI MODELS & PROVIDERS

10+ providers: OpenAI, Anthropic, Google, xAI, DeepSeek, Perplexity, Meta, Mistral, Qwen, Moonshot
40+ models across FREE, BASIC, and PRO tiers
Real-time streaming via SSE with token usage tracking and cost estimation
Multimodal input: text, images, and PDFs

AI SETS & MODEL RANKING

Curated model sets for quick switching between model combinations
Create, edit, delete, and cloud-sync sets with public directory and shareable links
Community-driven model ranking with voting, filtering, and quick vote mode

STORAGE & DATA

Local-first architecture with IndexedDB — conversations stay on your device by default
Optional PostgreSQL cloud backup for cross-device sync
Export conversations as HTML or Markdown
Global backup/restore (JSON/ZIP) with batch cloud backup

SHARING & COLLABORATION

Share conversations via public links with share history management
Email sharing with PDF attachments
Cross-conversation context inclusion for follow-ups
File browser panel with PDF and text file attachments

BILLING & SUBSCRIPTIONS

10 subscription tiers from Free to Enterprise
Monthly and annual billing with ~20% annual savings
Usage tracking with warning notifications and extra usage packs

USER EXPERIENCE

Authentication with guest mode for instant access
Profile settings with avatar upload
Usage history and monthly usage dashboard
Mobile-optimized with swipe gestures and responsive sidebars
Community feature voting page
Early access / waitlist onboarding flow

TECHNICAL

Frontend: React 19, Vite, TypeScript, Tailwind CSS 4, Zustand, Framer Motion
Backend: Node.js, Express 5, TypeScript, PostgreSQL
Auth: Clerk with JWT sessions
Deployment: Docker, Coolify, Hostinger KVM VPS
Security: Helmet, CORS, rate limiting, encrypted cloud storage