Releases

2026.3.9 — March 2026 (March 9, 2026)

A major release delivering AI-powered text messaging agents, team access controls, a platform administration dashboard, AI cost visibility, and dozens of improvements to messaging and contact management. Merges 50 PRs from dev into main.

Text Agents (new module)

AI-powered SMS/WhatsApp agents that auto-generate reply suggestions (docs)
Customization fields for tone, persona, and response style per agent
AI instruction builder with structured generation (docs)
Test conversation panel in the agent detail view
Phone number assignment and default channel agents (docs)

Authorization & Team Management

ABAC authorization across all 93 API endpoints with 15 resource types (docs)
Permission-aware UI that adapts based on user roles (docs)
Invite users to your organization via email (docs)
Pending invitations list with revocation

Super-Admin Platform Administration

Overview tab with platform KPIs, time-series charts, and org rankings (docs)
Organization management with health badges and drill-down detail views
System config with credential management and provider health verification
AI costs dashboard with per-model cost breakdowns

AI Cost Tracking

Per-organization usage logging across all 18 AI call sites (docs)
Realtime voice per-turn token tracking from OpenAI and Gemini sessions
Cost dashboard with per-model pricing table and cost-over-time charts (docs)

Messaging Improvements

Real-time updates via WebSocket — messages appear instantly (docs)
MMS file attachments with lightbox viewer (docs)
Contact info sidebar alongside conversations (docs)
Inline reply suggestions with auto-scroll (docs)
Smart reply routing from the phone number the contact originally texted
Markdown rendering in analysis, timeline, and messaging views

Contact Management

Communication feed replacing activity history
Contact enrichment from SMS history (docs)
Expanded contact schema with additional fields
Contact-user assignment with management UI
Note taxonomy with visibility, origin, and area classification (docs)
Note search in contact details
Last contacted tracking with aging indicators (docs)

Telephony & BYO Twilio

Redesigned BYO Twilio connect flow (docs)
SMS and call history sync from BYO accounts (docs, call sync)
Per-phone voice webhook configuration (docs)
One-click SMS event configuration for BYO accounts (docs)

Infrastructure

All AI services routed through OpenRouter (docs)
File storage migrated to Cloudflare R2 (docs)
Structured logging with Betterstack integration (docs)
Seed migration env fallback for automatic migrations on server startup

View on GitHub →

v0.1.0 — Initial Release (February 17, 2026)

The first tagged release of Loquent, capturing all work from project inception.

Features

AI voice agents with OpenAI Realtime & Gemini Live providers
Real-time transcription via GPT-4o Transcribe
Per-agent realtime configuration (model, voice, tools)
Knowledge base CRUD with LLM-powered queries
Contact management with auto-resolve
Post-call to-do extraction and AI execution
Per-call email reports with configurable toggles
System analyzers with default toggle
Sign up & phone number purchase
Agent CRUD, assignment, and management UI
Call history and details views
Scheduled daily reports (5 AM Eastern)
Styled HTML email reports
Authentication UI and backend
Configurable OpenAI & Gemini voice selection
OpenAI audio noise reduction
Speech interruption handling

Bug Fixes

Mono channel recording for reliable transcription
OpenAI realtime session cleanup on call end

Infrastructure

Clean architecture enforcement with panic safety
Graceful shutdown
Per-organization client lookup configuration
Core config table replacing environment variables
Project README, justfile, and .env.example

View on GitHub →