Hold review mode.
Checking combined auth, storage, and live Worker cutover status.
A living product roadmap for ChangEngine. Use this page to track what shipped, what is active, and what should come next as the coach companion becomes real.
Persist preparation, notes, maps, and experiments so coaching can continue over time.
Choose how Prepare, Session notes, map entries, and experiments are saved.
Account and persistence changes should move to production only after auth, journey storage, and CoachProfile storage are configured, smoked, and explicitly switched.
Checking combined auth, storage, and live Worker cutover status.
Checking access runtime readiness.
Waiting for readiness data.Checking auth activation preflight.
Waiting for readiness data.Checking journey storage readiness.
Waiting for readiness data.Checking database activation preflight.
Waiting for readiness data.Checking CoachProfile storage readiness.
Waiting for readiness data.Checking account settings storage readiness.
Waiting for readiness data.Checking Coach DNA source readiness.
Waiting for readiness data.Header gateway and Cloudflare Access are both supported. Keep review mode until one path has its variables, external provider, smoke command, and live request verified.
Waiting for /api/access/activation.
A practical queue for keeping ChangEngine moving in small, reviewable releases without losing the long-term Coach Engine vision.
The schema and repository adapters now cover journeys, account settings, and custom coach models; the next real unlock is connecting the managed database.
Configure DATABASE_URL, run migrations, seed data, smoke-test journey/account/custom-coach persistence, and switch all storage repository modes in deployment.
Storage status reports Postgres active for journeys, account settings, and custom coach models, and the strict journey API smoke proves live routes are using Postgres.
Cloudflare Access and signed-header adapter paths now exist, but production still needs a protected deployed provider configured with real users.
Protect the Worker with Cloudflare Access or an equivalent account layer, map real users into ChangEngine roles, and verify /api/access/status resolves non-review actors.
Production auth mode can distinguish client-owned journeys, coach-visible journeys, and admin review using real authenticated accounts instead of the review actor.
Growth Workshops can now be saved and reviewed, but Prepare, AI Chat, and live Session still need to open from the active workshop context.
Let a selected Growth Workshop seed Prepare, AI Chat, and live Session prompts with the current goal, protection, assumption, experiment, journal, and coach notes.
A coach can open a client journey, choose the current workshop, and start a session where the visible brief and AI context reflect that workshop instead of a blank journey.
Move ChangEngine from local single-device drafts into durable client journeys.
The workflow now has valuable local data that needs a real account-backed model before client use expands.
Acceptance: The schema covers users, journeys, sessions, prepare answers, notes, promoted map evidence, handoffs, Coach DNA sources, and experiments.Coaches need a clean artifact they can send, save, or review outside the live workspace.
Acceptance: A session handoff can be exported in a follow-up-ready format with map evidence and next experiment.Journeys, account settings, and custom coach models now have repository boundaries, but production still needs the managed database switch.
Acceptance: Credentialed Postgres validation passes and runtime storage modes can move from cache/static review mode into account-backed Postgres mode.Local persistence lets us test the Carl workflow immediately without blocking on auth, account models, and database setup.
The product should preserve agency and avoid presenting hypotheses as truth.
Showing the latest 8 of 149 shipped updates. The generated ROADMAP.md keeps the complete archive.
Added a ChangEngine-branded slide builder so a saved Growth Workshop can become a coach-ready slide and Google Slides-importable deck.
Add native Google Slides export after account-backed Google OAuth and Drive permissions are available for site users.
Built the first durable Growth Journey architecture so client workshops can be saved to Journey records, reviewed by coaches, assisted by AI suggestions, and summarized across workshops.
Connect the selected Growth Workshop into Prepare, AI Chat, and live Session so coaching starts from the active workshop context.
Added Growth Journey as the next major product layer so ChangEngine can organize transformation into workshops instead of only one-off preparation, chat, and session records.
Start by designing the Growth Workshop data model and a single Strategic Partner or Pricing From Value workshop before building the full category library.
Closed the gap between open coaching chat and the Journey record by letting users save the conversation and review transcript-derived map insights before a coach sees them.
Make the review step feel like a guided coach/client handoff: show what was captured, what still needs to be explored, and what the coach should review first.
Stopped AI Chat from treating each answer like an isolated prompt by sending recent conversation and current map context into the coach provider.
Add visible confirm/reject controls for AI-proposed map updates so context becomes a user-approved Change Map instead of passive chat history.
Made the selected Journey visible inside AI Chat so users can switch saved journeys without editing the URL.
Add a clear confirm/reject flow for AI-proposed change map updates so the user controls which suggestions become part of the Journey.
Moved AI Chat persistence from the demo-only snapshot to the same selected Journey context used by Prepare and Session.
Expose a visible Journey switcher in AI Chat so users can choose the saved Journey without editing the URL.
Connected AI Chat starter selection, transcript, stage, and edited change map fields to the existing Journey snapshot path.
Move the AI Chat route from the demo journey id to the currently selected saved Journey once the app-wide journey selector is active.
After every shipped development pass. The roadmap should show what changed, what evidence proves it, what moved status, and which three builds are now highest value.
Generate a release log draft from commits, then edit it while implementation details and verification evidence are still fresh.
Draft roadmap data changes in the private editor, then move item statuses based on what actually shipped.
Keep the build queue opinionated so the next session starts with a clear recommendation.
Treat the roadmap as part of the product surface, so it should be validated locally, checked in CI, deployed, and smoke-tested like any other page.
Create the first working ChangEngine surface and prove deployment.
Prove ChangEngine can ship as a real Cloudflare-hosted product, not just a concept.
App Router, OpenNext, Cloudflare Worker deployment, and live review URL.
Mock and OpenAI providers with structured output and safe runtime configuration.
Immunity to Change installed as the first pluggable methodology.
Settings-backed system, light, and dark appearance modes with local review persistence.
Settings now lets reviewers choose system, light, or dark appearance, with theme preferences included in local account data controls.
Cloudflare now uses the OpenAI provider with structured Responses API output.
The site is password-gated for private review during development.
Turn the leadership deck into a guided pre-session workbook.
Help a client arrive to a Carl session with useful self-observation instead of a blank page.
Personal Brand, desired brand, first ITC map, test run, and copyable session brief.
First Coach Engine profile with transparent inspired-by-Carl boundaries.
Mobile app bar for Home, Prepare, Session, and Roadmap access.
Let Prepare users open ChangEngine-styled previews of the leadership deck references while answering prompts.
Prepare now includes deck-source preview buttons and email-ready Carl session briefs.
Added a homepage-linked resource article explaining competing commitments, big assumptions, and small evidence-based tests.
PPT source material became a guided preparation workflow and copyable session brief.
Make ChangEngine useful during a real session with Carl or another coach.
Give Carl a calm, practical in-session companion that tracks the ITC map while he stays present.
Prep brief, coach notes, live ITC map, suggested coaching moves, and session handoff.
Carry completed Prepare answers into the Session workspace instead of showing sample content.
Let a coach promote, edit, and remove map evidence as hypothesis or confirmed entries with clear review history.
Generate a concise close-of-session summary with client language, hypotheses, open questions, and one next test.
Clarify the live coaching screen around one flow: review prep, capture notes, promote insight, and design one small test.
Make the right rail read as a closing checklist and collapse after-test reflection until it is relevant.
Show one map column at a time with compact navigation across Commitment, Behaviors, Competing commitment, Big assumption, and Test run.
Add a short Carl-style question and listening cue to each guided ITC map column.
Each guided ITC map column now includes a short coaching question and a listening cue for the active Session map step.
The Session map now works one ITC column at a time with compact navigation and editable evidence in the active column.
The Session right rail now starts with a simple now/during/close guide, and after-test reflection is collapsed until needed.
Save client work so sessions become an ongoing journey rather than one-off pages.
Persist preparation, notes, maps, and experiments so coaching can continue over time.
Prisma/PostgreSQL models for users, journeys, maps, notes, tests, and sessions.
Protect preparation, notes, map edits, experiments, and transition history with local and server journey autosave.
Debounce selected Session workspace changes into the journey API so server snapshots stay current without a manual Save click.
Generate coach/client summaries for follow-up after each session.
Shape one low-risk ITC experiment with an assumption, action, prediction, evidence plan, review date, and boundaries.
Capture what happened after a test, what changed, and the revised assumption without overwriting the original prediction.
Review planned tests, completed reflections, revised assumptions, and next steps across the journey.
Use sparse AI-generated audio playback beside vital Prepare and Session guidance without exposing OpenAI keys to the browser.
Prove signed actor headers offline or against a deployed header-mode environment before switching production out of review mode.
Prove live database save/load, coach visibility, shared-client visibility, and unrelated-client exclusion before activating Postgres.
Save Prepare answers into structured Postgres rows alongside the lossless journey snapshot.
Save current session notes and promoted ITC map evidence into structured Postgres rows alongside the lossless journey snapshot.
Save experiment timeline entries and reflections into structured Postgres rows alongside the lossless journey snapshot.
Save the exportable coach/client handoff summary into structured Postgres rows alongside the lossless journey snapshot.
Give clients a simple shared session room for notes, preparation review, and journey context while coaches open a focused cockpit from the same Journey record.
Let reviewers copy, download, and reset browser-local ChangEngine data while server-side account deletion waits for real auth and profiles.
Run a repeatable browser smoke for Prepare, Settings, and Roadmap at mobile widths before deploys.
Check visible mobile text line boxes so card-level clipping cannot hide Prepare hero or progress copy.
Run a repeatable accessibility smoke for landmarks, headings, named controls, labels, duplicate ids, and hidden focus traps.
Run a repeatable performance smoke for route timing, request count, transfer size, and HTML payload budgets.
Run a repeatable security smoke for review auth, security headers, safe API errors, auth activation state, and storage readiness.
Run one account-backed persistence activation command before and after DATABASE_URL is configured.
Run one production-auth activation command before and after a signed-header gateway is configured.
Run one combined production cutover gate before switching storage or auth runtime modes.
Show one combined production cutover verdict on the Roadmap and through an API.
Block Carl Mode source promotion until consented source records, review status, and revocation handling are persistent.
Prove authenticated journey create, save, restore, list, and membership paths before switching persistence adapters.
Require the journey API smoke to prove active Postgres adapter status during production storage cutover.
Resolve Cloudflare Access authenticated email headers into ChangEngine client, coach, and admin actors through server-side role allowlists.
Prove Cloudflare Access actor resolution offline and against a deployed Access-protected Worker before switching production auth mode.
Use Google and Apple identity providers for coach and client login so ChangEngine never stores user passwords or password reset secrets.
Let a coach sign in, create a client Journey, and send a client invite that opens the client's own Journey without exposing other coach records.
Advance journey snapshot versions when selected-journey preparation is saved so the persistence path has an auditable change trail.
Advance journey snapshot versions when selected-journey Session saves or autosaves update in-session state.
Reject stale Journey API writes so older client snapshots cannot overwrite newer server state.
Stream assistant text into the coaching conversation while preserving structured final metadata for safety, stage, and map updates.
Keep the live change map available while the chat scrolls and export the conversation by email or print-to-PDF.
Expose the same Email and PDF chat export controls on phone layouts where the desktop map rail is hidden.
Save the current chat to the selected Journey and review transcript-derived ITC map suggestions before confirming them.
Run one browser-backed review flow across Home, Prepare, Session, Settings, Roadmap, and mobile navigation before local and live releases.
Run one local command for focused logic, policy, repository, state, privacy, and preference tests without browser or external service setup.
Run ESLint through the release-quality gate and keep Prettier available for intentional formatting passes.
Remove the known production-build warnings so roadmap releases have a quieter quality signal.
Add the shadcn registry config, cn helper, first Button primitive, and contract test for future shared UI primitives.
Persist theme, audio, and privacy preferences through a static review adapter now and User.preferences in Postgres once authenticated accounts are active.
Expose authenticated account settings get/save routes and a Settings page sync bridge for theme, audio, and privacy preferences.
Require account settings storage readiness before production cutover can be marked complete.
Run account settings repository coverage inside the Postgres activation validator and document the paired repository-mode switch.
Verify account preferences save, normalize, reload, and clean up through User.preferences during credentialed Postgres activation.
Resolve conflicting dark-mode Prepare card, meter, pill, and disabled-action styling while preserving mobile layout guards.
Let users refine a preparation answer with AI from the field itself, with undo available before the original wording is lost.
Persist account-owned custom coach models through a CustomCoachModel table and Postgres repository once managed storage is active.
Expose the exact runtime variables, blockers, and validation command required before moving review storage and auth into production mode.
Show the managed DATABASE_URL, database smoke, repository switch, and live validation checklist directly in Settings.
Show signed-header and Cloudflare Access provider setup, blockers, smoke commands, and live validation directly in Settings.
Open the clean Presentation Page route with the latest in-page answers instead of stale static fallback data.
Keep local env, Cloudflare vars, docs, and validator handoffs aligned with the production storage switch plan.
Hide older saved journeys from the active index and restore them from Settings without deleting server snapshots.
Show a categorized review packet for local browser data that can later become account-backed after auth and Postgres activation.
Let reviewers choose which local data categories to export or import as a migration packet before account-backed import exists.
Expose signed-header and Cloudflare Access setup variables, smoke commands, blockers, and active states in the Roadmap readiness panel.
Accept selected migration packets on the server and apply preference data through the account settings repository while queueing journey data for Postgres import.
Convert valid local Prepare and Session migration packet data into repository-backed journey snapshots while keeping incomplete categories pending.
Normalize archived journey choices from selected migration packets so hidden journeys can become account-backed archive preferences later.
Summarize Coach DNA packet data for review without applying it until consent-audited source persistence and promotion are active.
Validate selected account migration packets through the server import boundary without saving preferences or journey snapshots.
Save and restore real journeys, settings, and custom coach models from the database instead of using localStorage or cache as the source of truth.
Session now defaults to a client-facing room with shared notes and preparation context, while coaches can open the coach cockpit from the same Journey using the coach role view.
AI Chat can now save the active conversation to the selected Journey and stage confirmable ITC map suggestions from the transcript for coach/client review.
Settings now has a production auth activation checklist with Cloudflare Access and signed-header provider paths, blockers, and smoke commands.
Use AI to assist the coach without taking over the coaching relationship.
Generate useful coaching support while preserving Carl's judgment, timing, and client trust.
Generate Carl-style coaching moves from notes and confirmed map state.
Add routing for crisis, therapy-like, or unsupported use cases.
Support future Coach Engines beyond Carl with consent-based profiles.
Review captured observations before they become approved Coach Engine profile material.
Validate stage, confidence, safety level, recommended action, and tool routing before AI suggestions affect state.
Route crisis, pause, careful coaching, and provider-failure states before normal coaching suggestions continue.
Convert provider state updates and classifier recommendations into confirmable map updates before mutating coaching state.
Route note promotion, experiment seeding, reflection promotion, and timeline capture through confirmable session transition proposals.
Save confirmed transition records with journey snapshots and coach-facing session exports.
Replay retryable provider-failure states from the last safe coach request without losing the user message.
Retryable provider failures now preserve the last safe request and show a Retry last message action in the coaching workspace.
Confirmed session transitions are now saved with journey snapshots, restored into Session, listed in the transition review panel, and included in session exports.
Session workspace actions now stage auditable transition proposals and require confirmation before promoting notes, seeding experiments, seeding reflections, promoting revised assumptions, or adding timeline entries.
Make audio feel intentional, configurable, and ethically clear as ChangEngine grows toward richer coach presence.
Let users hear key guidance without clutter while preserving consent boundaries for any coach-inspired or custom voice experience.
Allow users to choose a built-in OpenAI voice, playback preference, and speed for sparse audio cues.
Keep AI-generated audio disclosures visible and reusable across Prepare, Session, resources, and future coach surfaces.
Block custom coach voice use until consent recordings, source samples, and explicit usage permissions are configured.
Reviewers can now choose a built-in OpenAI voice and playback speed for sparse audio cues from /app/settings.
Sparse audio cues are now shipped in v0.4, with Settings, playback preferences, and consented custom coach audio tracked as planned v0.6 work.
Turn ChangEngine into a transformation platform where each client works through named ITC-based workshops over time.
Move beyond one map or one session by giving clients and coaches a workshop-based growth journey with context-aware AI, visible progress, experiments, reflections, and coach review.
Add the coach/client navigation model for Dashboard, Identity, Growth Journey, Workshops, Experiments, Journal, Insights, Coach, and Settings.
Let clients and coaches choose Professional, Financial, Spiritual, Personal, Relationship, Emotional, Legacy, or Custom as the container for each Growth Chart.
Remove preset templates from the visible build flow so each client starts with an empty category-backed Growth Chart and can add multiple charts in the same category over time.
Allow clients or coaches to create unlimited custom workshops while preserving the same ITC structure, experiment loop, and coach-review boundaries.
Represent each workshop as a durable record connected to a client Journey, with overview, current status, ITC map, experiments, reflections, journal entries, insights, coach notes, AI Coach context, progress, and session history.
Build one complete workshop, preferably Strategic Partner or Pricing From Value, before expanding into the full category library.
Route AI prompts through the active workshop context so the coach can draft ITC maps, suggest competing commitments and deeper assumptions, recommend experiments, summarize journal patterns, prepare sessions, and compare progress over time.
Track workshop movement through Awareness, Understanding, Experimentation, Habit Formation, Integration, and Transformation without implying guaranteed outcomes.
Let clients add journal entries and mark insights that attach to the active workshop and inform future AI and coach sessions.
Give coaches a focused mode to review all workshops, leave notes, suggest experiments, track progress, review journals, and prepare future sessions.
Let a workshop seed Prepare, AI Chat, and live Session context so coaching begins from the client's current growth track instead of a blank conversation.
Turn the active workshop map into a branded slide and download a Google Slides-importable PPTX for coaching sessions.
Create or update Google Slides directly from ChangEngine after account-backed Google OAuth and Drive permissions exist.
Keep language honest: progress represents observed engagement and confirmed learning, not diagnosis, therapy, or a promise of transformation.
Growth Journey now starts from selectable workshop categories, including Emotional Growth and Custom, and creates client-specific Growth Charts instead of presenting preset templates as work already started.
The Growth Workshop editor can now preview a ChangEngine-branded slide and download a Google Slides-importable PPTX.
Growth Workshops now save to Journey records, support custom workshop definitions, include journal and experiment history, offer AI-assisted draft suggestions, and provide cross-workshop coach insight rollups.