Skip to content

v1.22.0 — OpenRouter · AI Execution Modes · Copilot v2

Choose a tag to compare

View all tags
@qnbs qnbs released this 11 Jun 17:11
v1.22.0
This tag was signed with the committer’s verified signature.
qnbs qnbs
SSH Key Fingerprint: kAujKNM8X7UP/vh1MKsDL/bep+BUb39clB93nRIaC3c
Verified
Learn about vigilant mode.
56f567d
This commit was signed with the committer’s verified signature.
qnbs qnbs
SSH Key Fingerprint: kAujKNM8X7UP/vh1MKsDL/bep+BUb39clB93nRIaC3c
Verified
Learn about vigilant mode.

What's new in v1.22.0

✨ OpenRouter — Cloud 5 Free-Tier Gateway

Access 100+ open-source models through a single free API key:

  • Free models: DeepSeek R1, Llama 3.3 70B, Qwen 2.5 72B, Gemma 3 27B, Mistral 7B (:free suffix = zero cost)
  • Circuit breaker: 4 × 429 → 5 min pause, then auto-resume
  • Settings → OpenRouter: enable toggle, encrypted key storage, model selector
  • Sign up at openrouter.ai/keys — no credit card for :free models

⚙️ AI Execution Modes

Four routing strategies in Settings → AI & Models → AI Execution Mode:

Mode Behaviour
Hybrid (default) Local when preloaded → cloud fallback
Cloud All requests to configured cloud provider
Local On-device only — nothing leaves your device
Eco Battery-saving: tiny 0.5B model + heuristics only

AiModeIndicator chip in the Copilot header shows the active mode and turns amber when the OpenRouter circuit breaker is open.

✦ Global AI Copilot v2 — Complete

  • Markdown rendering — assistant replies render as formatted HTML
  • Sidebar mode — dock the panel to the right edge on desktop (≥ 768 px)
  • Apply to chapter — one-click rewrite of the active manuscript chapter (redux-undo reversible, Ctrl+Z)
  • InlineAnnotationLayer — insight-count badge inside the Manuscript Editor
  • ProForge Ask-Copilot chip — pre-fill the Copilot from any ProForge review item
  • Heuristics-only mode — brain-icon toggle for offline/zero-API analysis

🔧 Fixes

  • PWA blank screen on update: SW APP_VERSION kept at 1.20.0 caused the old cache to survive deployments. Bumped to 1.22.0 — existing PWA users receive an automatic cache-clearing reload after this update lands.
  • joi CVE (GHSA-q7cg-457f-vx79) resolved via lockfile patch

📊 Metrics

  • 2 594 i18n keys × 11 locales (en/de/fr/es/it + ar/he RTL Beta + ja/zh/pt/el Beta)
  • 5 475+ unit tests / 449 files
  • Coverage: lines ≥ 74 · branches ≥ 60 · functions ≥ 67 · statements ≥ 72

📖 New Docs

Full Changelog: https://github.com/qnbs/StoryCraft-Studio/blob/main/CHANGELOG.md

What's Changed

  • feat(proforge): audit & perfection cycle — apply edits, persistence, 8-stage supervisor, semantic RAG by @qnbs in #106
  • feat(proforge): dual-purpose Core Capability Layer + MCP server + Global AI Copilot by @qnbs in #107
  • fix(copilot): dismiss the assistant panel on outside click by @qnbs in #108
  • feat(ai,voice): WebLLM worker offload + Whisper STT E2E + voice hardening (P1-1/P1-2) by @qnbs in #109
  • feat(copilot): Ultimate Copilot AI Live Assistant — Phase 1 heuristics foundation by @qnbs in #110
  • feat(copilot): Phase 2 — markdown rendering, sidebar mode, Apply-to-chapter, InlineAnnotationLayer by @qnbs in #111

Full Changelog: v1.21.0...v1.22.0

Contributors

qnbs
Assets 10
Loading