Changelog
Platform updates, new features, and improvements.
New Feature
v3.2.0· Mar 9, 2026AI Email Assistant & Webhook Builder
Two powerful new tools to automate email workflows and event-driven architectures.
- •AI Email Assistant with context-aware drafting
- •Webhook Builder with multi-provider support
- •Brand Voice Trainer for consistent content
- •10 new blog posts and tutorials
New Feature
v3.1.0· Feb 28, 2026Multi-Agent Systems & Prompt Chaining
Orchestrate multiple AI agents and chain prompts into complex workflows.
- •Multi-agent team creation API
- •Prompt chaining with conditional logic
- •Agent communication protocols
- •Dynamic agent scaling based on task complexity
Launch
v3.0.0· Feb 15, 2026Platform V3: 400+ Models, Smart Router & More
Major platform upgrade with 400+ models, intelligent routing, and new media generation capabilities.
- •400+ AI models from 20+ providers
- •Smart Model Router with auto-selection
- •Video generation with Veo 3, Kling & Runway
- •Voice Studio with 100+ natural voices
- •Song Studio powered by Suno AI
Improvement
v2.8.0· Feb 1, 2026Developer Experience Upgrades
Faster APIs, better error messages, and new SDK features for Python and TypeScript.
- •50% faster API response times
- •Structured error responses with fix suggestions
- •TypeScript SDK v2 with full type safety
- •Batch processing concurrency increased to 100
New Feature
v2.7.0· Jan 20, 2026Fine-Tuning Marketplace & Model Versioning
Share and deploy community fine-tuned models. Pin model versions for reproducible outputs.
- •Community model marketplace
- •Model version pinning
- •One-click model deployment
- •Fine-tuning cost reduced by 40%
New Feature
v2.6.0· Jan 10, 2026Multi-Model Consensus Tools
New suite of tools leveraging multiple AI models for higher-confidence results.
- •Fact Checker with multi-model verification
- •Hallucination Detector
- •Debate Arena for model comparison
- •Contract Analyzer with risk scoring
Bug Fix
v2.5.1· Jan 5, 2026Stability & Performance Fixes
Critical fixes for API reliability and dashboard performance.
- •Fixed token counting for Claude models
- •Resolved rate limiting edge cases
- •Dashboard loading time reduced by 60%
- •Fixed BYOK key validation for Google AI