IER-0
PricingSign in
Home
Changelog

Product changelog

Product updates and milestone notes for the governed AI support cockpit, starting with the V1 baseline on May 20, 2026.

Start pilot View pricing
Latest
Jun 2

Newest entries reflect the latest committed main branch updates.

Window
May 20-Jun 2

Release notes start with the V1 baseline and follow the committed product history.

Focus
Safety

Auth, policy gates, auditability, and setup readiness.

1.0

June 2, 2026 - Support Supervisor & approved outbound follow-up call loop

Shipped the closed-loop voice follow-up workflow, featuring a Support Supervisor review queue, an approved outbound SIP call executor with quiet-hours gating, and strict human-in-the-loop voice controls.

  • Support Supervisor queue for AI-recommended voice call follow-up actions.
  • LiveKit outbound SIP call executor with automated quiet-hours (08:00-22:00 UTC) enforcement.
  • Human-gated execution loop with full audit trail coverage for approval decisions.
  • Standardized workspace support email fallback to support@driftrail.com.
2.0

June 1, 2026 - Branded support email editor

Added a project settings editor for the approved customer email shell, including brand identity, colors, signatures, footer text, preview, and test-send controls backed by audited server routes.

  • Per-project email branding state with a tenant-isolated migration.
  • Preview and test-send endpoints for the exact support email template.
  • Approved reply and auto-send paths now use the saved project branding.
3.0

June 1, 2026 - Founder Ops and launch stability pass

Tightened the launch control plane with clearer Founder Ops queue behavior, migration and health-check fixes, and a voice launch cleanup that keeps calls on the server-validated LiveKit/Cartesia path.

  • Founder Ops queue and repo-change PR flow polish.
  • Migration drift, health check, and CI E2E overlay fixes.
  • Removed the paused direct voice-cloning surface and covered the Cartesia catalog request shape.
4.0

May 29, 2026 - Read-only customer data lookup rollout

Added the public project data setup flow for read-only DB and signed HTTPS lookup connectors, with strict tool validation for safe customer-context SQL.

  • Per-project connector model with PostgreSQL URL and signed HTTPS relay.
  • Scope and schema allowlists for supported tables/views with blocked column patterns.
  • Email-bound lookup parameters (`$1`) and one-statement SQL restrictions.
  • Tool enablement for Agent drafting and widget surfaces with identity gating.
5.0

May 25, 2026 - Landing changelog access and team controls

Added a localized changelog path from the landing footer and continued the production-control pass for AI controls, voice routing, and approved knowledge context.

  • Localized landing footer changelog link
  • Customer-facing AI Controls view
  • Knowledge-backed voice context and SIP readiness guards
6.0

May 25, 2026 - Control plane hardening

Added admin control routes and scheduled maintenance tasks behind explicit access rules, plus schema-backed task logging so governance and approval evidence are explicit in dashboard workflows.

  • Admin console access and safe routing
  • Daily operations brief and cron orchestration
  • Task schema with approval-aware agent execution
7.0

May 25, 2026 - Auth and route hardening

Stabilized the authenticated entry flow, added visible Google OAuth actions, and fixed the protected-route CSP fallback so sign-in, sign-up, and guarded pages behave consistently in production.

  • Google OAuth entry actions
  • Stable auth slideshow routes
  • Protected route CSP fallback
8.0

May 25, 2026 - Dashboard access and send-policy safety

Hardened dashboard access and governed auto-send workflows so automation remains policy-bound, permission-checked, and described with user-facing copy that does not imply hidden model-owned side effects.

  • Dashboard permission checks
  • Auto-send policy documentation
  • Audit-backed send workflow updates
9.0

May 24, 2026 - Public readiness and localization

Shipped production-readiness guardrails, knowledge safety coverage, localized landing and dashboard routes, role permissions, production CSP asset allowances, and E2E stabilization for public-facing surfaces.

  • Readiness and knowledge safety tests
  • Localized dashboard and landing routes
  • Role permissions and production CSP assets
10.0

May 23, 2026 - Localization and setup polish

Expanded multilingual coverage across public and dashboard surfaces, refined project setup flows, improved DNS retry handling, and made configuration states easier to resolve from the dashboard.

  • Localized dashboard panels and marketing copy
  • Language switcher and dynamic AI prompt locale support
  • Clickable setup checks and clearer DNS status
11.0

May 22, 2026 - Production dashboard and channel readiness

Prepared the V1 dashboard for launch with governed auth screens, production data paths, onboarding and integration install polish, voice readiness, hosted phone provisioning, and database-backed E2E coverage.

  • Neon Auth launch hardening
  • Voice and phone-number readiness
  • Database-backed E2E stabilization
12.0

May 21, 2026 - App foundation

Initialized the application foundation and established the structure for the support cockpit, documentation set, and V1 implementation plan.

  • Next.js app baseline
  • Support cockpit documentation structure
  • Initial implementation path
13.0

May 20, 2026 - V1 product baseline

Defined the governed-support baseline: workspace-scoped operations, policy-gated AI outputs, human approval for side effects, auditability, and the first complete V1 product surface.

  • Security-first product contract
  • Model proposes, policy decides, executors act
  • V1 dashboard and channel scope
Operations

What this means in production.

These pages describe the product contract behind the UI, not a decorative brochure. Each surface should connect back to the same governed support loop.

Control boundary

Entries reflect committed work where repository history exists, with May 20 used as the V1 baseline marker.

Operator workflow

Support moves through secure ingest, AI preparation, human approval, outbound execution, and audit evidence.

Configuration impact

Provider setup, project policies, knowledge trust, retention, and billing limits decide how this behaves for a real workspace.

Proof points

Designed for governed support, not hidden autonomy.

Entries reflect committed work where repository history exists, with May 20 used as the V1 baseline marker.
Automation language stays aligned with the V1 rule: models propose, policy and humans decide, executors act.
Production-readiness claims are tied to committed hardening, documentation, and test coverage.
References

Official stack and legal references.

These external links point to provider documentation, privacy notices, data-processing terms, or platform terms that shape production operation.

Vercel DPAexternalHosting, Functions, Blob, and Workflow data-processing terms.Neon Auth Next.js SDKexternalOfficial auth SDK reference used for app sessions and identities.Google Gemini APIexternalDirect Gemini model and structured-output API documentation.
Next

Keep exploring the operating model.

Operating signalsWorkspace settingsPricingSign in