Public roadmap

Dashloom is developed in small, verifiable product slices. Dates are intentionally omitted until the preceding milestone passes its completion checks.

Milestone 1 — Product foundation

• [x] Public English product page
• [x] Simplified Chinese product entry
• [x] Multi-workspace D1 schema
• [x] Product, connector, mapping, metric, and sync-run entities
• [x] Canonical, hreflang sitemap, robots, and social metadata
• [x] CI, dependency updates, security and contribution policies

Milestone 2 — First real connection

• [x] Authentication and workspace onboarding
• [x] Product management in the browser
• [x] Multiple Cloudflare account connections
• [x] Workers Analytics resource discovery with manual-name fallback
• [x] Idempotent manual synchronization
• [x] Connector health and useful error recovery
• [x] Unified multi-account lifecycle with explicit health, credential rotation, local credential erasure, mapping shutdown, and audit events
• [x] Product-level mapping, stored-evidence, freshness, and specialist-Agent coverage diagnostics
• [x] Evidence-backed activation path from first product to recurring Agent report
• [x] In-product feedback and strict anonymous diagnostic export
• [x] Guided real-product ingestion with product-scoped write keys and evidence/Agent readiness verification

Milestone 3 — Google and reliable analytics

• [x] Google OAuth with PKCE
• [x] Multiple Google accounts
• [x] GA4 and Search Console resource discovery
• [x] Product-to-resource mapping
• [x] Scheduled sync, compare-and-claim execution, retry backoff, and plan limits
• [x] Deterministic period comparisons, anomaly detection, and evidence bundles
• [x] Deterministic product health scoring with Agent-visible provenance

Milestone 4 — AI product intelligence

• [x] Five dashboard presets: Indie Hacker, SaaS Revenue, SEO Growth, Cloudflare Operations, and Agency Client
• [x] Workspace-scoped agent profiles and permission boundaries
• [x] Structured Agent Playbooks for business model, objective, priorities, sensitivity, output style, and language, frozen into every run
• [x] Product operating goals with deterministic rolling-period progress and Agent-citable target-gap evidence
• [x] Deterministic Agent Signal Radar for material metric movement, health risk, goal gaps, Growth Mission drift, cross-signal context, and explicit one-click specialist investigation
• [x] Bring-your-own OpenAI-compatible endpoint, API key, and model
• [x] Evidence-linked conversational analysis
• [x] Evidence-linked Agent Action Center with deduplicated occurrences, assignment, due dates, lifecycle, backfill, and audit
• [x] Post-completion Agent outcome loop with frozen metric baselines, later-evidence measurement, recurring cycles, and non-causal disclosure
• [x] Agent Growth Missions with operator-approved targets, frozen metric identity and baseline, automatic progress measurement, assignment, audit, and Agent-citable learning evidence
• [x] Persistent workspace-scoped conversation threads with bounded, non-evidentiary history
• [x] Server-locked portfolio or real-product conversation scopes with scoped readiness, evidence minimization, and Signal Radar handoff
• [x] Server-locked portfolio or real-product scope across manual reports, proactive alert explanations, recurring schedules, Executive Briefs, and delivered history
• [x] Portfolio, revenue, SEO, operations, and client-reporting agent skills
• [x] Prompt-injection boundaries, bounded evidence, citation validation, quota, and contract tests
• [x] Five-role golden dataset, deterministic quality rubric, and replaceable output evaluator
• [x] Deterministic competitor-period trends and non-causal cross-signal evidence
• [x] Workspace Agent Quality Lab with identical frozen evidence, prompt-versioned provider results, and deterministic comparison statistics
• [x] Executive Brief orchestration across two to five independent specialists with failure isolation, deterministic prioritization, and original-run provenance
• [x] Scheduled daily, weekly, and monthly Executive Brief reports with cadence-matched evidence, occurrence idempotency, retries, and channel delivery
• [ ] Publish captured live cross-provider benchmark results with model versions and parameters

Milestone 5 — Reports and managed Dashloom Cloud

• [x] Registration and email-relay account recovery
• [x] Team invitations and role-based access
• [x] Plan-based managed AI allowances and append-only usage ledger
• [x] Daily, weekly, and monthly report workflows
• [x] Operable report schedules with pause/resume/delete, execution health, stable failure codes, and bounded retry backoff
• [x] Email relay, Slack, Discord, and webhook delivery with retries
• [x] Deterministic threshold and percentage-change alerts with cooldowns, optional evidence-linked Agent explanations, failure-safe fallback, and auditable delivery retries
• [x] Usage quotas and Worker observability
• [x] Idempotent Cron execution ledger, stale-run recovery, and operator-visible task outcomes
• [x] Plan-aware data retention controls and daily cleanup
• [x] Stripe-hosted subscription billing, signed webhooks, reconciliation, and entitlements
• [x] One-time Studio trial eligibility, usage limits, cancellation/refund request tracking, and verified operator queue
• [x] Live deployment reachability status and public support-response policy
• [x] Workspace export, deletion, and audit
• [x] Portable Community-to-Cloud product and evidence import with privilege stripping
• [x] Customer-safe workspace export and D1 backup/recovery runbook
• [ ] Recurring independent backup restore drills in the hosted production environment

Milestone 6 — Ecosystem

• [x] Stripe restricted-key, Lemon Squeezy, privacy-minimized Creem, and least-privilege Polar revenue connectors
• [x] Paddle Billing as an additional merchant connector beyond Stripe, Lemon Squeezy, Creem, and Polar
• [x] GitHub repository activity connector
• [x] Vercel deployment health connector
• [x] Supabase project health and aggregated service-request connector
• [x] Cloudflare R2 operations and storage analytics with explicit provider-retention evidence
• [x] Privacy-minimized Cloudflare Pages deployment evidence
• [x] Cloudflare Queues best-effort realtime backlog and paused-delivery evidence
• [x] Connector SDK and authenticated normalized ingestion API
• [x] Customer-side Cloudflare Connector Worker with D1 binding, fixed aggregate contract, Cron delivery, and secret-only ingestion authentication
• [x] Safe scheduled Custom REST pull connector with a versioned normalized metric contract
• [x] Competitor evidence import and Agent-visible comparisons
• [x] Agent Skill manifests and bounded runner integration
• [x] Product-scoped, expiring and revocable client dashboards
• [x] Agency client workspaces and private deployment support
• [x] Agent Skill install-time policy validation, immutable version upgrades, frozen instruction fingerprints, and automated contract evaluations
• [x] In-product public Connector/Skill catalog with server-resolved installation and maintainer-review provenance
• [x] Community submission intake with schema/policy CI and current-head independent GitHub approval verification
• [ ] Complete and publish the first independently reviewed community extension
• [x] Reusable dashboard templates with custom metric and product scope
• [x] Agent-generated smart dashboards with specialist mapping, live metrics, and frozen-analysis provenance
• [x] Expiring, revocable, product/metric-scoped read-only dashboard sharing
• [x] Deterministic calculated metrics with product/date/currency alignment
• [x] Workspace white-label branding across the authenticated product, reports, and client portals

Feature requests are welcome, but new work should reinforce the core promise: turn trustworthy product signals into evidence-linked explanations, recommended actions, and recurring reports.