Changelog

What's New

Track the latest updates, new features, and bug fixes in SpecKit.

v1.8.0

October 2, 2026

Latest

Added

• In-App Notification System: Dashboard bell notifications with a detail modal to read admin feedback responses and system updates directly.
• What's New / Changelog Widget: Dashboard release announcement popover shown once per user, fully manageable via Admin Panel (/admin/changelog).
• Inline Admin Feedback Reply: Ability for admins to reply to user feedback inline directly within the Admin Feedback dashboard.

Fixed

• API Security Hardening & Input Sanitization: ObjectId validation, strict length limits on feedback/changelog inputs, and public field sanitization across API endpoints.

v1.7.0

September 30, 2026

Past

Added

• Shadcn UI Combobox Component: Upgraded native HTML select dropdowns in the Blueprint Review Step to searchable Shadcn Comboboxes built on Radix UI (Popover & Command) for smooth tech stack selection.
• Health Endpoint Automated Regression Test Suite: Added automated Vitest testing for /api/health to prevent regression of system environment metrics.

Fixed

• Sanitised Information Leakage in /api/health Endpoint: Fixed sensitive server diagnostic data leaks (uptime & node environment) while preserving HTTP 200 'ok' status and Uptime Kuma health monitor compatibility.

Changed

• Workspace Tech Stack Presets Modernization: Updated default tech stack recommendations to latest modern framework versions (Next.js 15+ App Router, React 19, Tailwind CSS v4, Prisma v6, Node.js 22, FastAPI Python 3.12+, Astro 5, NestJS 11).

v1.6.0

September 27, 2026

Past

Added

• 100% Dynamic On-the-Fly AI Interview Engine: Real-time question generation tailored to project context with a 5-question limit and automated branch pruning when previous answers are updated.
• Custom Tech Stack Selection: Freedom to add custom frameworks, databases, and libraries beyond predefined default options.
• SDD Cross-Document Consistency & Blueprint Export Auto-Reconciliation: Automated integrity validation across specifications, with export tolerance that automatically resolves missing referenced API endpoints (e.g., from USER_FLOW.md) into API.md on-the-fly during ZIP creation.

Fixed

• Hardened architecture prompts across all AI providers (Vertex AI, DeepSeek, Bedrock) to guarantee the generation of two mandatory diagram types: graph TD and sequenceDiagram within ARCHITECTURE.md.
• Harmonized AI Provider method signatures (Vertex AI, DeepSeek, Bedrock) and resolved mock interface type errors.

v1.5.6

September 21, 2026

Past

Added

• Expanded SEO and AI visibility with topic-based documentation, software planning guides, spec-driven development comparison pages, and crawlable educational content.
• Added centralized sitemap data, route-specific canonical metadata, JSON-LD structured data, expanded robots.txt rules, and llms.txt for search and AI crawlers.

Changed

• Updated initial HTML to use English consistently for crawlers, with internal linking and keyword targeting for PRD, requirements, planning, and spec-driven development.

v1.5.5

September 19, 2026

Past

Added

• AI-Powered Dynamic Specification & Interview Engine: Recommended document suites (PRD, API, ERD, Security, Compliance) and interview questions now scale dynamically based on project complexity and ambiguity without artificial static caps.
• Expanded Idea Input Ceiling up to 20,000 Characters: Allows users to paste long-form architecture specs or project briefs across the Workspace Wizard and Editor.

Fixed

• Security Patch for Project Repurposing (Payload Hijacking): Hardened /api/ai/generate with project ID ownership checks and database payload overrides to prevent quota bypass exploits.
• Patched Infinite Generation Flaw: Enforced mandatory Blueprint quota checks for all initial workspace document creation routes.

Changed

• Tier-Based Interview Question Scaling: Free tier accounts receive up to 5 focused interview questions, while Starter & Pro users unlock expanded 7-8 question deep dives for ambiguous project inputs.

v1.5.4

September 18, 2026

Past

Fixed

• Security hardening for Polar Webhook route (/api/webhooks/polar), eliminating default fallback to 'starter' to prevent entitlement corruption from invalid payloads.
• Implemented strict Pro plan downgrade protection to prevent active Pro subscriptions from reverting to Starter on incomplete or unrecognized webhook events.
• Refined user identity resolution prioritizing metadata and verified subscription history to prevent arbitrary user updates from non-unique email fallbacks.
• Integrated atomic transaction claiming and automatic 5-minute stale lock recovery for Polar webhooks to guarantee idempotency and eliminate race conditions.

v1.5.3

September 16, 2026

Past

Fixed

• Automated handling for Polar subscription lifecycle events (`subscription.uncanceled`, `subscription.active`, and `subscription.updated`) ensuring instant plan restoration upon reactivation.
• Refined Polar subscription cancellation logic (Cancel at Period End), allowing users to retain their paid quota and workspace access through the end of their active prepaid billing cycle.
• Added intelligent fallback mechanisms in Polar webhooks for resolving user IDs and subscription plans.

v1.5.2

September 3, 2026

Past

Fixed

• Patched authorization guard in Admin Settings (/api/admin/settings) by properly awaiting admin authentication verification.
• Hardened AI document generation & chat routes (/api/ai/generate & /api/ai/chat) with strict userId ownership validation to prevent premium paywall bypass.
• Eliminated webhook idempotency race conditions in Midtrans & Polar using atomic MongoDB state transitions (findOneAndUpdate) to prevent duplicate token grants.

Changed

• Upgraded core security dependencies: Next.js (v16.2.11), Mermaid.js (v11.16.1), and DOMPurify (v3.4.13).

v1.5.1

September 3, 2026

Past

Changed

• Updated Terms of Service Section 4: Enhanced clarity regarding Polar & Stripe payment integrations, subscription cancellation workflows, and immediate quota reset policies upon reverting to the Free Plan.

v1.5.0

August 19, 2026

Past

Added

• Interactive Workspace Simulator component on the landing page featuring custom markdown rendering, live Mermaid diagram visuals, and real-time interactive drafting.
• Comprehensive Admin Dashboard (/admin) with real-time user activity tracking, AI token cost estimations, and system maintenance toggles.
• Dual-Column Vertical Marquee Testimonials section with infinite smooth scrolling (left up, right down) and refined minimalist card aesthetics.

Fixed

• Complete fix for Export ZIP Blueprint Suite: forced .md file extensions, 3-layer data resolution (State, SessionStorage, & MongoDB), and total elimination of empty dummy files.

Changed

• Recalibrated AI max output token ceilings (PRD.md expanded up to 20,480 tokens / 20K, technical specs up to 7,500 tokens) to prevent truncated responses.
• Refactored modular AI Provider architecture with full support for DeepSeek V3/V4, Google Vertex AI (Gemini 2.5), and AWS Bedrock (Claude 3.5).

v1.4.0

August 13, 2026

Past

Added

• Bilingual AI Documentation Pipeline: Target language selector (Indonesian & English) at the workspace summary screen before document generation.
• Mobile-responsive Language Preference card with adaptive layout (vertical stack on mobile, sm:grid-cols-3 on desktop).
• Dynamic language prompting across AI providers enforcing strict English for technical syntax (Prisma ORM, Mermaid ERD, API routes).

Changed

• Recalibrated forging specs visual progress countdown timer to 105 seconds for PRD and 70 seconds per technical document batch.

v1.3.3

July 27, 2026

Past

Added

• Dynamic database-driven document counters inside the Blueprint Suite card, showing real document counts directly from the database and optimized using Next.js unstable_cache with a 5-minute revalidation window and a smooth requestAnimationFrame count-up animation.

Fixed

• Aligned testimonials Bento Grid cards to sit perfectly straight on desktop while maintaining mobile-responsive sticky stacking animations.

Changed

• Redesigned the Hero Section simulation card footer into a vertical stacking layout to eliminate text overlapping and layout shifts across all devices.

v1.3.2

July 26, 2026

Past

Added

• Double Subscription Prevention: Automatically disabled active or downgrade-eligible plan purchase buttons on the pricing page, backed by server-side checkout validation on Midtrans and Polar APIs.

Changed

• Modernization of Quota Ledger Audit Logs: Redesigned the admin ledger table into a clean monochrome aesthetic, standardizing columns (Type, Description, Date, Amount) and dynamically classifying entry types.

v1.3.1

July 25, 2026

Past

Fixed

• Token quota loss when project creation fails due to database errors (integrated automatic token refund mechanism)
• Inconsistent loading state ('isGenerating') in document workspace after successful AI specification generation
• Dashboard project history skeleton loader disappearance during network or API data loading failures

v1.3.0

July 25, 2026

Past

Added

• Full AI multi-provider integration (AWS Bedrock, DeepSeek, Vertex AI) with auto failover and strict usage quotas
• Interactive Mermaid.js diagram viewer with zoom and pan capabilities in the editor page
• Strict 'STRICT SCOPE' guardrails in chat to block cross-document content generation (e.g. database ERD/APIs in unrelated specs)
• Automatic redirect (unauthorized project redirect) to workspace dashboard with alert toast on illegal/cross-account project access attempts

Fixed

• Blank editor loading state on unauthorized/non-existent project IDs by capturing and handling HTTP 403/401/404 responses

Changed

• Enhanced document content length and complexity constraints tailored for Free, PAYG, Starter, and Pro plans inside LLM system prompts
• Migrated middleware.ts to proxy.ts to align with Next.js 16 requirements, resolving development API routing issues (404/settings)

v1.2.0

July 23, 2026

Past

Added

• Tiered monetization system with dynamically gated software specifications (Free, PAYG, Starter, Pro)
• Interactive glassmorphic Lock Overlay for premium documents in workspace
• One-token unlock system with predefined bundles (e.g. Tech Spec, DevOps, QA) or custom selections (up to 3 files)
• Pro-tier exclusive SEO & Search engine optimization strategy document generator (supporting llms.txt, JSON-LD schemas, and Edge middleware)
• Dozzle logging service for real-time container log monitoring in docker-compose configuration
• Automated fallback error states in editor with manual 'Retry Generation' button to prevent token loss on connection issues

Fixed

• Token race condition protection at API level with automatic refunds on parallel concurrent requests

Changed

• Updated Create Wizard to bypass locked documents during initial project generation, replacing them with a preview lock placeholder

v1.1.6

July 16, 2026

Past

Added

• Dynamic quantity selector (1-5) for purchasing Pay-as-you-go (Eceran) blueprint tokens
• Dynamic ad-hoc pricing override system integrated with Polar checkout gateway

Changed

• Hardened Polar & Midtrans webhooks against metadata parameter manipulation exploits (anti-spoofing)
• Repositioned the quantity dropdown directly above the CTA button on the pricing cards for better visibility
• Adjusted quota remaining language structure for Indonesian locale to read more naturally ('sisa 9' instead of '9 tersisa')

v1.1.5

July 15, 2026

Past

Added

• Interactive workspace generator screen with dynamic estimation countdown
• Labor illusion live processing logs and educational SpecKit tips carousel
• Flexible zoom and pan capabilities for rendered Mermaid diagrams
• Automated failover and fallback mechanism to a secondary provider when the primary server encounters rate limits

Fixed

• Enhanced Mermaid diagram parsing and character sanitization for robust rendering

Changed

• Updated generation screen progress bar to a solid black design matching the site identity
• Migrated generation UI typography to Geist Sans
• Improved search engine optimization (robots.txt) for protected user paths
• Consolidated assistant chat and document impact analysis into a single, more efficient API call
• Smart prompt classification to show or hide document sync buttons based on query intent

v1.1.0

July 13, 2026

Past

Added

• High-performance multi-engine AI integration with token consumption tracking
• Automated failover and fallback systems to ensure uninterrupted document generation service availability
• Admin overview dashboard with real-time server statistics and revenue tracking
• User retention cohort heatmap dashboard
• Admin feedback management panel and database integrity audit tools

v1.0.0

July 2, 2026

Past

Added

• Initial public launch
• AI-powered project understanding and smart interview
• Tech stack recommendation engine
• Documentation generator with 4-15+ adaptive docs
• Blueprint workspace with three-panel layout
• AI assistant with impact analysis
• Version history and rollback
• ZIP export for all documentation
• Pricing plans: Free, Pay-as-you-go, Starter, Pro
• Bilingual support (English & Indonesian)

v0.9.0

June 20, 2026

Past

Added

• Beta testing program
• Markdown preview with syntax highlighting
• Document search functionality

Fixed

• Database schema generator timeout issues
• PDF export formatting on large documents
• Interview question loop edge cases

Changed

• Improved AI response accuracy for complex architectures
• Updated pricing page with currency toggle (USD/IDR)
• Redesigned landing page with forge animation

v0.8.0

June 1, 2026

Past

Added

• Tech stack customization modal
• Real-time document generation preview
• FAQ page with bilingual support

Fixed

• Auth redirect loop after sign-out
• Workspace editor scroll jump on long documents

Changed

• Migration from OpenAI to Google Gemini for cost efficiency
• Workspace UI polish with Material Symbols icons

This changelog is updated with every new release.