The everything-you-touch-first release. Five weeks tightening creation, play clarity, and the reliability of your saved stories.

What's new

• Creation, rebuilt. The character/ship/crew wizard is now one clear question per screen on a unified design system, with a "Play now" fast start (a ready-made character, straight into your first mission) alongside build-your-own. Focuses and Values are browsable by discipline and moral theme with search and description cards, and every step's writing was rewritten in Wanderhold's own voice across all five worlds.
• The rules, in reach. A new Rules reference at /rules, an in-wizard rules drawer, and inline term tooltips throughout play — tap any game term (Momentum, Threat, a discipline, a roll) for a plain explanation without leaving the scene.
• A better home for your campaigns. Redesigned world selection and campaign cards with per-world artwork; saved campaigns show mission progress and resume cleanly across devices; account settings unified into one place (Campaigns, Preferences, Plan & Billing, Profile).
• Missions that read like stories. Every mission gets its real AI-written title again, pacing guarantees each mission a genuine climax, and complications name a concrete mechanism instead of vague atmosphere.

Reliability & fixes

• Your progress is safe. A rewritten single-transaction save path with version guarding, fixes for silent cloud-save data loss, dependable cross-device resume, and automatic recovery for stranded campaigns.
• The story keeps moving. Bounded AI retry/backoff, higher token limits that end mid-campaign scene failures, and removal of an infinite-retry loop.
• Resume just works. Returning to a campaign you saved before starting now generates your first mission automatically instead of dead-ending on an empty screen.
• A large mobile and world-fidelity sweep — overflow, tap targets, popover clipping, per-world vocabulary, WCAG-safe theming — plus analytics and signup-attribution wiring behind the scenes.

Post-launch polish wave. A week of work tightening the parts of the game players touch first: a new world, ready-to-play characters, campaigns that remember themselves, and a long sweep of creation and play-screen fixes.

What's new

• A fifth world in draft state: Hammerward. A warm stone-and-castle setting joins Stellar Odyssey, Wind & Plunder, The Territories, and Erebus Station, with its own theme, tables, and conveyance.
• Pregenerated characters. Every world now ships with ready-made characters you can pick and play immediately — skip the full build when you just want to start a story.
• Campaigns that remember. Threads you leave open in one mission now carry into the next, recurring NPCs come back by name, and your ship's name and timeline stay separately pinned across missions. The back half of a campaign holds together.
• Character creation overhaul. Always-visible option descriptions, a redesigned "Set Sail" review, auto-generated and editable campaign names, a cleaner Ship's Habit step, a mobile-first creation flow, and dozens of smaller copy and layout fixes.
• Reworked play header. Identity, act, and tabs are reorganized into a clearer always-on header with act-progress dots and a sticky tab seam.

Fixes & polish

• Journal export renders more reliably, with an expanded table of contents and better per-mission grouping.
• Complication resolve now writes a narrative beat instead of silently clearing, and injuries show a "Treated" confirmation when healed.
• Lighter, more legible theming — light-theme color tokens, a canonical type scale, and Frontier accent colors corrected to pass WCAG AA on parchment.
• Smoother first paint on the marketing site and a fix for the crew-stats step crash during character creation.
• Upgrade nudges that make sense — players at the campaign limit are routed to pricing with a clear banner instead of a dead end.

Public launch. Wanderhold goes live with the full solo-RPG engine, four worlds, billing, and the polish we wanted in players' hands on day one.

What's new

• Status Bar v4 — the new 4-zone heads-up display surfaces momentum, threat, complications, and feedback in one always-visible row. Designed for at-a-glance reads during a scene.
• Decompression-aware narration. Scenes now distinguish "the engine is setting something up" from "the engine is closing it." The result: fewer flat openings, fewer rushed endings, more weight where it belongs.
• Mission ending polish. Upgrade success and account-tier flows got a top-to-bottom redesign so finishing a campaign or upgrading a tier feels like a moment, not a form submission.
• Public release notes. This page. Bumps will be called out here going forward.

Tooling and tone. Two weeks of work focused on observability for me and texture for you.

What's new

• Game Log / Ledger drawer. Every mechanical beat in a scene — rolls, spends, complications — is now auditable inline. Open the drawer to see exactly why something resolved the way it did.
• Narrative FX. Card crawl, breathing, and staggered reveals on key story beats. Small touches; the scene feels less like a wall of text.
• Mobile status panel redesign. Tabbed slide-up, Lucide icons, compression-aware tokens. Status reads cleanly on a phone now.
• Blog v2. /blog ships with an editorial workflow: write markdown, sync to Supabase, publish. Five drafts queued.
• /forge admin pipeline stepper (admin-only) — a 20-stage step-through visualizer of how the engine generates a mission, from oracle rolls through prompt construction. Seedable for reproducibility.
• Custom-action test mode + 10 player-personality presets for replaying scenes with different decision profiles in the audit pipeline.

Core-loop quality wave. A focused sprint on the moment-to-moment storytelling itself.

What's new

• Chekhov identity, Act 3 pacing, campaign hooks — three coordinated changes to make the back half of a campaign land instead of fizzle. Threads opened in early missions now resolve with weight at the end.
• Test/prod prompt parity. The prompts we score in the lab are now the prompts the production engine sends, removing a category of drift bugs.
• /worlds hub page — a single browsable index of every world, with parallel marketing copy and improved per-world UI.
• Location persistence. The engine now extracts and remembers where the current scene is happening, surfaces it as a read-only chip, and threads it through subsequent prompts.
• Frontier temporal eras. The Frontier world spans 1865-1890 and now respects the era you picked at campaign start.

Hardening sprint + mobile pass. Quieter under-the-hood work that mostly shows up as "fewer things go wrong."

What's new

• Architecture hardening (Phases A-E). Error boundaries, stricter ESLint module boundaries, reducer migration for mission state, discriminated game-state phases, cross-world contract tests. Crashes are dramatically less common.
• Mobile UX overhaul — tab overflow, grid stacking, button alignment, and contrast fixed across the play screen. Mobile is genuinely playable now.
• SEO + bot verification. 14 marketing pages now pre-render so search engines and OAuth bots see real HTML instead of a blank shell.
• Onboarding polish + admin gate. Cleaner step bar, world pre-selection, era descriptions; /editor and other admin tools properly gated.

Billing goes live. Subscriptions, tiers, and the staging gate to keep tire-kickers off pre-production builds.

What's new

• Stripe billing with three tiers: Wanderer (free), Saga, and Chronicle. Server-side price IDs, webhook signing, upgrade success page.
• Era-aware terminology. All hardcoded "Captain / phaser / ion storm" engine vocabulary is now resolved through a per-world term system so each world speaks its own language end-to-end.
• Staging password gate keeps casual visitors off the staging URL while letting agents and Playwright in via a shortcut.
• Staging-first deploy pipeline. Production now only deploys after a clean staging build, with CI gates on both.

Worlds go live. Erebus Station and Frontier flip from draft to playable; the multi-world theme system is the visible difference.

What's new

• Erebus Station + Frontier live as fully playable worlds alongside Stellar Odyssey and Wind & Plunder.
• World theme design system — each world gets its own palette, typography accents, and chrome that swap at runtime via [data-theme].
• Blueprint archetype system. Missions now build from a 10-archetype blueprint that gives each mission a structural promise the engine commits to keeping.
• Mission briefing UI. Pre-mission scene leads with stakes and the opening choice instead of buried setup.
• World isolation fixes + analytics instrumentation. No more cross-world leakage of names or vocabulary.

Two new worlds in draft. Erebus Station and Frontier are added as world data with their own tables, conveyance systems, and themes.

What's new

• Erebus Station (Antarctic mystery) — full world data and a default/arctic theme pair. Slow-burn isolation atmosphere.
• Frontier (western) — full world data, conveyance system, light-background theme. Draft status; goes live in 0.5.
• Play Quality Sprint. AI timeout handling, error recovery, spend hints, "Continue the Story" affordance, PostHog Surveys dual-control wiring.
• World Editor improvements — empty-table state, breadcrumb, progress bars, AI table generation, comparison view.

Wanderhold name + Layer 1 design system. The platform formally takes the Wanderhold name and rebuilds the public site on a new visual system.

What's new

• Wanderhold rename + domain migration to playwanderhold.com.
• Layer 1 (Steel + Ember) design system applied across every marketing page.
• Legal page suite — Privacy, Terms, Safety, Refunds, Accessibility.
• Account page with profile management and tier visibility.
• Vitest test suite migration — 93 test files, 1,766 assertions, 0 failures. The legacy runner moves to non-blocking.
• AI Playwright QA agent — Claude-driven exploratory browser testing with real Supabase auth, separate from CI.

Auth, analytics, and stakes. Players can sign in; the engine can track them; the mechanics get teeth.

What's new

• Magic-link sign-in via Supabase — PrivateRoute, AuthScreen, callback handling, sign-out flow.
• PostHog analytics wired with P0 activation events (campaign created, scene generated, mission completed, etc.).
• Crew + character hit trackers for extended-conflict scenes — opposition damage finally has structure.
• AI prompt audit pass — broad sweep on quality and tone of generated scenes.
• Organic complication injection — complications now arrive from the world, not pasted on after the fact.

Engine foundations. The first version where the loop holds end-to-end: pick a world, build a character, run a mission, save, resume.

What's new

• CSS theme system — palette swap at runtime; engine reads colors from CSS variables, not hardcoded.
• URL routing for setup wizard, play tabs, and the dashboard.
• World Editor MVP with admin tools.
• CI/CD pipeline via GitHub Actions.
• Sentry observability including AI latency spans and a weekly digest.
• Pirates world brought to deep parity with the original world.
• API connectivity test suite covering Supabase and Anthropic.

Engine prototype before public-facing work began. Stellar Odyssey mechanics running in a single hardcoded world, no auth, no marketing site, no billing. Everything from 0.1.0 onward is built on this base.