Changelog

What's new.

Every feature, fix, and improvement — documented as it ships.

v0.11.0April 20, 2026

Phase 2 is now fully shipped: graph-based Prompt Chains, safer MCP tools, richer artifacts, and better model comparison telemetry.

FeaturePrompt Chains now execute as graphs instead of a simple list — add branching paths, choose an entry step, and guard against loops during runs

FeatureChain editor now supports step keys, explicit success/false routing, entry-step selection, and chain import/export

FixDuplicating or importing a chain now remaps IDs correctly so graph links stay intact

FeatureMCP tools now infer risk level, require approval for higher-risk actions, and keep a local audit log of approvals/executions

FeatureArtifact previews now support SVG and Markdown in addition to HTML / React / Mermaid

ImprovementArtifacts can now be edited inline and downloaded directly from the preview panel

ImprovementModel comparison now surfaces duration, token counts, cost metrics, and exportable snapshots for each run

ImprovementAll four recent implementation slices now pass typecheck clean and production build verification

v0.10.0April 20, 2026

Phase 2 foundation release: Graph view, installable PWA, App Lock, encrypted .novakit backups, auto-archive, and BYOS sync.

FeaturePrompt Chains now have a read-mostly Graph tab — visualize the success + branch edges of your workflows with dagre auto-layout

FeatureInstallable PWA: NovaKit works offline via a service worker with an /offline fallback and an offline banner

FeatureApp Lock: optionally require your vault passphrase to open the whole app, with an auto-lock timer

FeaturePanic Wipe: one-click erase of every trace of NovaKit from this browser — IndexedDB, localStorage, caches, and the service worker

FeatureEncrypted .novakit backup files (AES-256-GCM, passphrase-derived key) with merge or replace import

FeatureAuto-archive conversations after a configurable threshold (30 / 60 / 90 / 180 days). Pinned chats stay put.

FeatureStorage breakdown card in Settings → Storage: per-table row counts and approximate bytes

FeatureBYOS Sync: bring-your-own-storage encrypted cross-device sync via WebDAV, fully implemented.

FeatureGoogle Drive and Dropbox sync use your own OAuth credentials — NovaKit never ships developer client IDs

FeatureContext Isolation toggle added to chat — optionally start fresh context when switching providers mid-conversation

DocsNew blog posts: Phase 2 ship notes, user-owned OAuth pattern, and 'NovaKit works offline now'

DocsFAQ expanded with entries on App Lock vs vault lock, Panic Wipe, encrypted backups, and offline mode

v0.9.0April 15, 2026

Privacy and trust upgrades, interactive savings calculator, and PDF support in the knowledge base.

FeatureInteractive savings calculator so you can estimate your costs vs a flat subscription

FeaturePDF uploads now supported in the knowledge base

FeatureBrowser storage is now persistent — your data won't be silently evicted

FeatureNew Storage page in Settings showing usage breakdown and quota

FeatureExport All Data: download a full backup of your workspace as JSON

FeatureAnthropic and Mistral now work seamlessly in the browser without extra setup

ImprovementSwitching providers mid-conversation now shows a clear privacy notice

ImprovementFirst knowledge base upload explains how embeddings are processed

ImprovementLicense settings now explain exactly what data is shared during activation

DocsNew FAQ entries covering telemetry, tracking, and phone-home behavior

DocsPrivacy policy expanded with per-tier data handling and GDPR rights

DocsPublic Roadmap and Changelog pages

v0.8.0April 15, 2026

Full marketing site with blog, about, contact, FAQ, and legal pages.

FeatureAbout page with our mission, values, and how NovaKit works

FeatureContact page with a simple form (no account needed)

FeatureBlog with rich posts, tags, and categories

FeatureThree launch blog posts: BYOK savings, privacy architecture, and model comparison guide

FeatureStandalone FAQ page organized by topic

DocsPrivacy Policy, Terms of Service, and Cookie Policy

ImprovementRSS feed for blog subscribers

ImprovementSearch engine crawlers blocked from private app routes

v0.7.0April 15, 2026

Live model pricing, device-based licensing, and updated free tier limits.

FeatureModel pricing now syncs from live data instead of a static list

FeatureDevice limits per plan: Free gets 1, Starter gets 2, Pro gets 5

FeatureFree tier now has a 20-message-per-conversation limit

ImprovementFree tier adjusted to 5 conversations for a better upgrade path

ImprovementCost calculator and model picker now use live pricing data

v0.6.0April 14, 2026

Knowledge base for document-powered conversations, plus license activation and tier-gated features.

FeatureKnowledge base: upload documents, organize into collections, and search them in conversations

FeatureDocuments are automatically chunked and embedded for intelligent retrieval

FeatureAI answers can now reference your uploaded documents for context-aware responses

FeatureLicense key activation with one-click tier upgrade

FeatureFree tier limits with clear upgrade prompts when you hit them

ImprovementEncryption passphrase is now optional — you can add it later in Security settings

v0.5.0April 14, 2026

Cost tracking dashboard, prompt library, personas, quick actions, and productivity features.

FeatureCost dashboard with daily spend charts and per-model breakdowns

FeatureBudget alerts that warn you when monthly spend approaches your limit

FeaturePrompt library: save, organize, and reuse your best prompts with template variables

FeaturePersona system: pre-built AI personalities for coding, writing, research, and more

FeatureCommand palette (Ctrl+K) for quick navigation and actions

FeatureQuick actions on any response: Summarize, Translate, Simplify, Draft Email, Action Items, Fact Check

FeatureFocus mode: hide the sidebar for distraction-free conversations

FeatureKeyboard shortcuts for common actions

FeatureOnboarding wizard to help new users set up API keys

FeatureImport conversations from ChatGPT and Claude

v0.4.0April 14, 2026

Expanded to 7+ AI providers, file uploads, and an improved model selector.

FeatureAdded Google Gemini, Mistral, Groq, Ollama, and OpenRouter

FeatureConnect any OpenAI-compatible endpoint (LM Studio, vLLM, etc.)

FeatureUpload images, text files, and code files to conversations

FeaturePaste images directly from your clipboard

FeatureModel selector with provider grouping, search, and favorites

v0.3.0April 14, 2026

Full conversation management with folders, tags, search, and export.

FeatureSidebar with conversation list and multiple view modes

FeatureDrag-and-drop folders to organize conversations

FeatureColor-coded tags for quick categorization

FeaturePin important conversations and archive old ones

FeatureSearch across all your conversations instantly

FeatureExport any conversation as Markdown or JSON

v0.2.0April 14, 2026

Core chat experience with OpenAI and Anthropic, real-time streaming, and cost tracking.

FeatureChat with OpenAI models (GPT-4o, GPT-4o-mini, etc.) with real-time streaming

FeatureChat with Anthropic models (Claude Sonnet, Opus, Haiku)

FeatureRich markdown rendering with syntax-highlighted code blocks

FeatureOne-click copy on code blocks

FeatureStop generation mid-stream

FeatureEdit your messages and regenerate AI responses

FeaturePer-message token count and cost display

v0.1.0April 14, 2026

Initial release: encrypted local storage, API key management, and the foundation for everything that follows.

FeatureAll your data stored locally in your browser — nothing on our servers

FeatureAPI key management with strong encryption

FeatureDark and light theme with system preference detection