IMS-Toucan Releases

v0.19.0 adds advanced permissions with Row Level Security (RLS), finalizes Deno support, and introduces mode helpers to simplify configuration and deployment.

Oct 31, 2025

v0.18.1 focuses on bug fixes, specifically addressing pagination issues when total counts are unavailable and adding documentation for Cloudflare Image Optimization.

Oct 13, 2025

v0.18.0 introduces three execution modes (UI-only, code-only, hybrid) and admin customizations, while changing the configuration property name from `initialConfig` to `config`.

Oct 1, 2025

This release focuses on stabilizing build and runtime environments, specifically fixing Cloudflare serving and raw node execution issues. It introduces a new UUID validation utility and enhances the PostgreSQL connection API. Dependencies were also bumped to version 0.17.1 to align with the latest improvements.

Sep 15, 2025

This release adds a Korean Resident Registration Number recognizer with checksum validation and integrates Azure Health Data Services (AHDS) for healthcare de-identification.

Sep 9, 2025

This release adds extensive support for reasoning models like DeepSeek R1 and Gemini 2.0 Flash-thinking, alongside dynamic model management for OpenAI and Anthropic providers. It also introduces new UI features like diff-view v3, persistent settings, and Netlify one-click deployment.

Feb 28, 2025

This release expands provider compatibility by integrating AWS Bedrock models (Claude 3, Nova, Mistral) and adding a GitHub provider. It also improves Git import functionality and Docker configuration.

Jan 22, 2025

A minor hotfix release that resolves an issue preventing the auto-selection of starter templates.

Dec 31, 2024

Focuses on error handling and UI improvements, including automatic code template detection and enhanced terminal alerts.

Dec 31, 2024

This release introduces a new GUI that provides precise control over how utterances sound, allowing users to generate different realizations, manipulate pitch and duration values by dragging, and exchange voices while preserving intonation and duration changes.

Oct 7, 2024

A minor release with major impact on audio quality, this update trains the vocoder on synthetic spectrograms instead of ground-truth audio, closing the quality gap and greatly improving output while maintaining fast inference speed.

Sep 22, 2024

This release provides new checkpoints with stochastic prosody prediction that samples from distributions, additional IPA modifier support, more languages in pretraining, and overhauled language similarity prediction modules with visualization.

Jul 25, 2024

A major release providing TTS support for almost all 7000+ languages in the ISO-639-3 standard through extrapolation from a pretrained 462-language checkpoint, with watermarking to prevent misuse and overall quality improvements.

Jun 10, 2024

This release enables emotional TTS by allowing users to condition the model on emotional prompts during training and transfer emotions from any prompt to synthesized speech during inference.

Jun 10, 2024

Introduces ToucanTTS, a new architecture designed for multilingual and low-resource research, accompanied by stable pretrained models and a new BigVGAN vocoder option.

Apr 10, 2023

This release is the official submission for the Blizzard Challenge 2023.

Apr 4, 2023

Enhances controllable multilingual capabilities by introducing a 24kHz vocoder, a flow-based postnet, and improved speaker generation controls.

Feb 22, 2023

Adds controllable speaker features through self-contained embeddings and replaces the vocoder with Avocodo for better performance.

Oct 25, 2022

Expands language support to include all IPA phonemes, suprasegmental markers, and tonal languages like Chinese and Vietnamese.

May 20, 2022

Enables multi-language and multi-speaker synthesis with a self-contained aligner and joint modeling of linguistic and speaker features.

Mar 1, 2022