ArcanFlowsBack to Home
Last updated: December 2025

Changelog

Track our journey as we build the future of AI-powered automation. See what's new, improved, and fixed in ArcanFlows.

v0.9.0Major ReleaseDecember 2025

Unified LLM Infrastructure

Complete overhaul of LLM backend management with multi-provider support, high availability routing, and GPU monitoring.

New Features
  • -Multi-backend LLM routing with automatic failover
  • -Support for Ollama, vLLM, TGI, MLX, and SGLang backends
  • -Real-time GPU monitoring dashboard with utilization, memory, and temperature metrics
  • -Model equivalence groups for HA fallback across backends
  • -Priority-based request queuing per backend
  • -Alert system with Log, Webhook, and Slack channels
Infrastructure
  • -New LLM Infrastructure management UI (/llm)
  • -Backend health monitoring with automatic status updates
  • -Model registry with cross-backend visibility
  • -OpenAI-compatible chat completions API endpoint
v0.8.0Major ReleaseDecember 2025

Embeddable Widgets

Enable forms and AI chat agents to be embedded in external websites with customizable themes and positions.

New Features
  • -Iframe embeds for forms and agents
  • -Chat bubble widgets (floating buttons like Intercom/Drift)
  • -JavaScript widget loader for easy integration
  • -Theme support (light/dark/auto)
  • -Position configuration (bottom-right/bottom-left)
  • -PostMessage API for iframe-parent communication
Improvements
  • -Embed code generator UI with live preview
  • -Copy-to-clipboard for embed codes
  • -React component code snippets
v0.7.5Minor ReleaseDecember 2025

Session Management & Memory Persistence

Enhanced session handling with expiration warnings and improved conversation memory persistence.

New Features
  • -Session expiration warning modal with countdown timer
  • -Options to extend session or logout gracefully
  • -Automatic token refresh functionality
Bug Fixes
  • -Fixed conversation memory loss after 24 hours
  • -PostgreSQL fallback when Redis cache expires
  • -Automatic Redis rehydration from database
Improvements
  • -Increased max_tokens validation limit from 32K to 1M
  • -Support for modern LLM context windows
v0.7.0Minor ReleaseDecember 2025

Enhanced Web Search

Improved web search reliability for local LLM models with intelligent query detection.

New Features
  • -Smart search intent classification (FORCE, TEMPORAL, VERIFY, GENERAL)
  • -40+ explicit search patterns in English, Spanish, and Portuguese
  • -Domain-specific keywords (stock prices, cryptocurrency, exchange rates)
Improvements
  • -Case-insensitive JSON query detection
  • -Enhanced system prompts for local models
  • -Structured logging for search decisions
v0.6.0Major ReleaseDecember 2025

External Chat Interface

Public-facing chat interface for agents with streaming responses and history persistence.

New Features
  • -Public chat page at /chat/{agent_id}
  • -Real-time streaming responses with SSE
  • -Conversation history with auto-scroll
  • -Message persistence in PostgreSQL
  • -Rate limiting per IP and agent
Security
  • -Agent must be published to be accessible
  • -Configurable rate limits per agent
  • -IP-based request throttling
v0.5.0Major ReleaseNovember 2025

Multi-Factor Authentication

Enhanced security with TOTP-based two-factor authentication and recovery codes.

Security
  • -TOTP-based 2FA with QR code setup
  • -Recovery codes for account recovery
  • -MFA status tracking per user
  • -Secure key storage with encryption
New Features
  • -MFA settings page in user profile
  • -QR code generation for authenticator apps
  • -Recovery code display and regeneration
v0.4.0Major ReleaseNovember 2025

Tools & Integrations System

Flexible framework for connecting agents to external services, APIs, and databases.

New Features
  • -REST API, Function, Webhook, Database, and Script tool types
  • -OAuth2, API Key, Bearer, and Basic authentication support
  • -11 predefined tool categories (CRM, Email, Analytics, etc.)
  • -Tool execution audit logging
  • -Agent-tool many-to-many relationship
Infrastructure
  • -JSONB configuration for flexible tool settings
  • -JSON Schema validation for inputs/outputs
  • -Usage tracking with automatic counters
v0.3.0Major ReleaseNovember 2025

RAG & Knowledge Base

Document processing and retrieval-augmented generation for informed AI responses.

New Features
  • -Document upload and processing (PDF, TXT, DOCX, HTML, MD)
  • -Vector embeddings with sentence-transformers
  • -Semantic search with ChromaDB
  • -Knowledge base management per agent
  • -Chunking strategies and overlap configuration
Performance
  • -Batch processing for large documents
  • -Background job queue for embeddings
  • -Caching layer for frequent queries
v0.2.0Major ReleaseOctober 2025

Form Builder & Automation

Visual form builder with AI-powered field suggestions and automated workflows.

New Features
  • -Drag-and-drop form builder
  • -15+ field types (text, select, date, file, etc.)
  • -Conditional logic and field dependencies
  • -Form templates and cloning
  • -Response collection and export (CSV, JSON)
Improvements
  • -AI-powered field suggestions
  • -Form preview mode
  • -Mobile-responsive form rendering
v0.1.0Major ReleaseSeptember 2025

Initial Release

Core platform with agent creation, multi-tenant architecture, and conversation management.

New Features
  • -Agent creation and configuration
  • -Multi-tenant architecture with role-based access
  • -Conversation management with context retention
  • -Model provider support (OpenAI, Anthropic, Ollama)
  • -Real-time chat interface
Infrastructure
  • -PostgreSQL with TimescaleDB for metrics
  • -Redis for caching and pub/sub
  • -Docker Compose deployment
  • -JWT authentication

Ready to get started?

Join thousands of users who are already building intelligent workflows with ArcanFlows.

Start Free TrialView Documentation