Skip to content

v1.9.353

Latest
Latest

Choose a tag to compare

View all tags
@michaelbeijer michaelbeijer released this 13 Mar 17:03
· 41 commits to main since this release
v1.9.353
9dc27ef

Highlights

AI Assistant Now in the Right Panel (v1.9.353)

The Supervertaler Assistant is now available as a tab in the right-side panel, right next to the Match Panel. This means you can chat with the AI while working in the translation grid — no more switching to a separate tab. The Prompt Manager and Variables tabs remain under the AI tab for configuration tasks.

All Changes Since v1.9.342

New Features & Improvements

  • AI Assistant moved to the right panel — visible alongside the grid for instant access while translating (v1.9.353)
  • Ollama streaming support — AI Assistant responses now stream incrementally instead of waiting for the complete response, eliminating timeout issues for large requests (v1.9.352)
  • Auto-generated prompts saved to dedicated folder — "Analyze Project & Generate Prompts" now saves to a "Supervertaler Assistant Prompts" folder (v1.9.351)
  • Multi-file support for standalone SDLXLIFF import — importing 2+ SDLXLIFF files now activates the full multi-file system with file boundary banners, file filter dropdown, and Manage Views dialog (v1.9.349)
  • File-aware AI prompt generation — detects multi-file projects and runs per-file domain/tone analysis (v1.9.349)
  • Expert-level AI prompt generation — domain-specific templates (patent, legal, medical, technical, financial, marketing) with anti-hallucination controls, terminology consistency hierarchy, and preflight self-checks (v1.9.347)
  • Simplified glossary priority system — replaced multi-level numeric priorities with a simple Project/Background toggle (v1.9.343)
  • Horizontal tabs for Resources — saves screen space for the 4 resource sub-tabs (v1.9.345)

Bug Fixes

  • Fixed Ollama timeout for AI Assistant and large prompts — regex-based model size detection and automatic timeout boost for large prompts (v1.9.352)
  • Fixed HTTP proxy not applied to most LLM calls — all 13+ LLM call sites now correctly pass the proxy setting (v1.9.351)
  • Fixed "Ollama is not running" false alarm — endpoint URL sanitization prevents double-path URLs like /api/api/tags (v1.9.350)
  • Fixed AI prompt generation stripping its own content — skip cleaning for prompt generation requests (v1.9.349)
  • Fixed AI prompt generation failure — switched from fragile JSON format to delimiter-based output (v1.9.348)
  • Fixed chat line break rendering — phase status lines now display on separate lines (v1.9.348)
  • Fixed Alt+Up/Down quick-add glossary shortcuts — explicit activation check before reading priority (v1.9.344)
  • Removed per-term priority column — completed glossary simplification, priority determined solely at glossary level (v1.9.346)

Install

  • pip: pip install supervertaler==1.9.353
  • Windows standalone: Download the .zip below, extract anywhere, and run Supervertaler.exe
  • macOS standalone: .dmg will be attached shortly

First-time Windows setup

  1. Download and extract the .zip to a folder of your choice (e.g. C:\Supervertaler)
  2. Run Supervertaler.exe
  3. Optionally run create_start_menu_shortcut.ps1 to add a Start Menu entry

Upgrading

Replace your existing Supervertaler folder with the new extracted folder. Your projects, settings, and termbases are stored separately and will be preserved.

Requirements

  • Windows: Windows 10 or later (64-bit)
  • macOS: macOS 12 Monterey or later (Intel & Apple Silicon)
  • Python (pip install): Python 3.10+

Links

Assets 4
Loading