Changelog

What's new in Konvoq

Every update, improvement, and fix documented as it ships.

Feb 20, 2026
v2.4.0
New Feature

Multi-LLM Routing & A/B Testing

You can now run multiple LLM backends simultaneously and split traffic between them to compare quality, latency, and cost before committing to a model.

Added LLM router: route by topic, user segment, or traffic split percentage
A/B test dashboard shows side-by-side quality and cost metrics for each model
Support for Claude 3.5 Sonnet added alongside GPT-4o and Gemini 1.5 Pro
Automatic fallback to secondary model on provider timeout or rate limit
Model-level spend caps to prevent budget surprises
Feb 7, 2026
v2.3.2
Bug Fix

Widget Rendering & Safari Fixes

Addressed a batch of widget rendering issues reported by users on Safari 17 and iOS 18, plus several edge cases in the conversation state machine.

Fixed chat widget flicker on Safari 17 caused by backdrop-filter repaints
Resolved iOS keyboard pushing widget off-screen on mobile
Fixed conversation context occasionally being dropped after 20+ message turns
Corrected avatar image crop on high-DPI displays
Jan 22, 2026
v2.3.0
New Feature

Advanced Analytics Dashboard

The analytics suite has been completely rebuilt. You now have access to conversation heatmaps, resolution funnels, revenue attribution, and exportable reports.

Conversation heatmap: see engagement peaks by hour, day, and week
Resolution funnel: track how many chats resolve vs. escalate vs. abandon
Revenue attribution: connect resolved chats to downstream purchases via UTM params
Top unanswered questions report updated daily
CSV and PDF export for all dashboard panels
Jan 8, 2026
v2.2.1
Performance

Response Latency Improvements

Infrastructure and prompt optimizations that cut median first-token latency by 38% across all regions.

Switched to streaming completions by default for all LLM providers
Knowledge base lookup now uses approximate nearest-neighbor indexing (HNSW)
CDN edge caching for static widget assets load time down from 420ms to 90ms
Reduced average context window size with smarter chunk ranking
Dec 18, 2025
v2.2.0
New Feature

Team Collaboration & Shared Inbox

Pro and Business plan users can now invite team members, collaborate on escalated tickets, and manage human handoffs from a unified inbox.

Shared inbox: all escalated conversations in one place for your team
Internal notes: leave private comments on any conversation
@mentions: notify teammates directly from the conversation thread
Role-based access: Admin, Agent, and Viewer permissions
Audit log: full history of who did what and when
Nov 30, 2025
v2.1.0
Improvement

Widget Customization v2

The widget editor got a major overhaul with live preview, new layout options, and far more granular control over every visual element.

Live preview panel see changes in real-time as you edit
Custom CSS injection for advanced brand consistency
New launcher button styles: circle, pill, and custom image
Conversation starter bubbles: up to 6 quick-reply prompts
Multilingual welcome messages based on browser locale
Oct 15, 2025
v2.0.0
New Feature

Konvoq 2.0 Complete Rebuild

After months of work, Konvoq 2.0 ships with a new training engine, a redesigned dashboard, and the foundation for everything we're building next.

New AI Training Engine with 3x faster ingestion and higher answer accuracy
Completely redesigned dashboard faster, cleaner, more powerful
Webhook system for real-time event notifications to any endpoint
Public API v2 with full OpenAPI spec and SDK libraries
New pricing plans with higher limits and more flexibility
Sep 2, 2025
v1.9.4
Security

Security Hardening & SOC2 Prep

A focused release on security hardening as we complete our SOC2 Type II audit, plus GDPR data residency controls for EU customers.

EU data residency option all data stored in Frankfurt region
Session tokens now rotate on every API call (token refresh pattern)
Added rate limiting to all public-facing API endpoints
Dependency audit: updated 34 packages with known CVEs