SpecLens v1.2.0

• Spec checks (beta) - SpecLens now lints changes and specs locally and deterministically: missing documents, malformed EARS/Gherkin blocks, empty or duplicated requirements, RFC 2119 misuse, ambiguous wording. Findings render as IDE-style underlines with hover explanations, plus a new Checks view in the sidebar (group by change or by check type), a resizable results panel scoped to what you're reading, and severity badges across the app. On by default; Settings → General can disable it or opt archived changes in
• Repositories with large git histories load dramatically faster: authorship now comes from a single history walk instead of one per file (a 4,500-commit repo with 2,800 specs went from hanging indefinitely to about 4 seconds)
• Only one right panel (comments, spec checks, AI summary) stays open at a time
• Updated dependencies: Vite 8, TypeScript 7, React plugin 6, js-yaml 5

SpecLens v1.1.3

• Fixed the Linux AppImage failing to start on Wayland with EGL_BAD_PARAMETER: the app now preloads the system Wayland client library instead of the bundled one and disables unstable WebKitGTK rendering paths inside AppImages (#13)

SpecLens v1.1.2

• SpecLens now tells you when a newer version is available: a notice on launch, a badge on the About icon, and a link in the About dialog. One request to GitHub per day, and it can be turned off in Settings → General
• Settings reorganized into General / Reading / AI tabs
• The onboarding tutorial covers the AI summary feature
• Changelog version numbers link to the code diff between releases

SpecLens v1.1.1

• Fixed a crash when quitting the app after using AI summaries (Metal resources are now released before exit)

SpecLens v1.1.0

Local AI - on-device summaries, fully private. Nothing runs or downloads until you enable it and fetch a model; nothing ever leaves your machine.

Highlights

• AI summaries everywhere - a golden ✨ button on the project Overview and on every document (proposals, tasks, design docs, specs, ADRs, playbooks) streams a summary into a resizable right-side panel. Capability names link straight to their specs.
• Reviewer-oriented document summaries - what the document is, its key points, and a "worth a reviewer's attention" list of risks and open questions.
• Background generation - keep reading while it works; a notification tells you when the summary is ready. Summaries cache per document and regenerate only when content changes.
• Five curated models - Gemma 4 E2B (default) and E4B, Qwen3.5 4B, Phi-4 Mini, SmolLM3 3B - downloaded on demand with checksum verification, all managed from Settings.
• Bring your own model - import any llama.cpp-compatible GGUF file, or drop it into the models folder.
• Ollama support - models from a local Ollama server appear automatically and generate through it.
• Settings redesign - a scannable model list with per-row download/delete, storage totals, and a simpler two-tab layout.

Known limitations

• AI inference is Metal-accelerated on Apple Silicon and CPU-based elsewhere; small models can be slow on older machines
• Windows installers remain unsigned - SmartScreen will warn on first run

SpecLens v1.0.3

• Fixed a crash when opening a document containing a Mermaid diagram before the diagram renderer had loaded ("Something went wrong rendering this view")
• Git is no longer required on your machine: authorship and timestamps now come from an embedded git library, with identical rename-following and .mailmap behavior

SpecLens v1.0.2

• Faster startup - the eager bundle shrank from 1.9 MB to 744 KB; Mermaid now loads on demand, only when a document renders a diagram
• The project went public: website, Homebrew install, CONTRIBUTING.md, and unit tests running in CI

SpecLens v1.0.1

• Add-repository picker now validates the selection - folders without an openspec/ directory are rejected with a clear warning instead of being persisted as broken sources
• Install through Homebrew: brew install --cask dansreis/tap/speclens (Apple Silicon), with the cask updated automatically on every release
• README: feature screenshot gallery, demo video, and a pointer to the openspec-examples try-it-now dataset

SpecLens v1.0.0

First release of SpecLens - a desktop reader for OpenSpec projects.

Highlights

• Browse OpenSpec repositories - add any local folder containing an openspec/ directory; proposals, tasks, and specs render as rich markdown with GFM, Mermaid diagrams, and EARS keyword highlighting
• Seven views - Overview, Changes, Specs, Flow, Graph, Timeline, and Schemas
• Git-derived attribution - created/edited bylines and timestamps come straight from repo history, with a per-repo cache for fast cold starts
• Comments - select any passage to attach a comment; unresolved/resolved workflow, quote-click jumps back to the highlighted text, persisted locally in SQLite; export as structured markdown ready to paste into an LLM conversation
• Navigate quickly - global search (⌘K), back/forward history (⌘[ / ⌘]), document zoom, and a minimap with a slide-out table of contents
• Signed and notarized macOS builds - installs and opens with no Gatekeeper warnings
• Onboarding tutorial on first launch, replayable from Settings
• About dialog - version, links, and these release notes, from the info button in the sidebar
• Customization - dark/light theme (the native window chrome follows it), drag-resizable sidebar, reading speed, highlight color, and comments panel width
• Multi-platform builds - macOS (Apple Silicon .dmg), Linux x64/arm64 (.AppImage, .deb, .rpm), Windows x64 (.msi, NSIS) and arm64 (NSIS)

Known limitations

• Windows installers are not signed yet - SmartScreen will warn on first run