Changelog

Relayer

The latest features, improvements and fixes, newest first.

Subscribe via RSS
  1. Update

    v1.3.0 Provenance and a redesigned serving sheet

    Every release now carries its provenance, and the serving sheet got the redesign it needed: facts, controls and lifecycle actions each have their own place.

    Added

    • Source column on the releases list: see at a glance whether a release came from GitHub, the API or the dashboard. Existing releases are classified retroactively.
    • Relayer's own changelog now lives on the landing page, rendered live from /changelog/relayer.

    Changed

    • Release sheet rebuilt with a clear hierarchy: read-only Details (status, source, channel, created, offers), Adoption, Serving controls (targeting, rollout) and Lifecycle (pause, resume, roll back) with the consequence of each action stated next to its button.
    • Import from GitHub dialog is wider, so release titles and dates no longer truncate.

    Docs: https://www.relayercli.com/docs/concepts

  2. Update

    v1.2.0 - Entry activity and editor polish

    changelog

    Every changelog entry now shows its life story, and the editor stops guessing about your intent.

    Added

    • Activity timeline on every changelog entry: who created it and how (GitHub webhook auto-draft, import, AI generation, or a person), plus every edit, publish and status change since - sourced from the audit log, so it cannot drift from reality

    Changed

    • The save button is dirty-aware: disabled until something actually changes, then comes alive; label reads "Save draft" only while the entry is a draft, "Save changes" once published
    • Imported entries now carry their GitHub release provenance (release tag and repo) instead of looking hand-written

    Fixed

    • No-op saves no longer fire API calls or pollute the entry's activity with phantom "Edited" events
    • Activity panel no longer refetches on every keystroke in the editor
    • Save button no longer shifts layout when toggling between enabled and disabled
  3. Update

    v1.1.0 - Targeting, rollback fallback, explainable rollouts

    rolloutstargetingapi

    v1.1.0 · Targeting, rollback fallback, explainable rollouts

    Serving grows up: choose WHICH devices get a release, know exactly what happens to devices on a pulled build, and ask the system why any device gets any version.

    Added

    • Label targeting per release (tier=canary), fail-closed with fall-through; l.* params on /u/, inline labels on decisions
    • Rollback fallback: stranded devices are served the newest published release below as a downgrade (allowDowngrade), with stranded counts and fallback preview before you commit
    • "Why this update?" decision trace + per-device Next-check line - the same walk production serves
    • Engineering blog at /blog with three launch posts
    • Manage serving from CI: PATCH /api/v1/releases (rollout, pause, rollback, targeting)

    Changed

    • Rollout % and targeting edit freely until the first offer, then only move forward
    • Releases tab redesigned: data table, channel filters, serving side panel
  4. Update

    v1.0.0 · GA - GitHub connect + the changelog product

    githubchangelogai

    v1.0.0 · GA - GitHub connect + the changelog product

    Relayer 1.0: connect GitHub and your changelog writes itself from what you actually merged - then you edit, publish, and it IS your public changelog.

    Added

    • GitHub App connect + per-app Source tab (PRs, commits, releases, tags)
    • AI changelog generation: checkpoint-based "since last publish" or per-release scope, PR-link toggle
    • Rich editor (TipTap), categories, tags, featured images, editable publish dates
    • SEO entry detail pages, curated changelog RSS, sitemap coverage
    • Webhook auto-draft on every GitHub release; import past releases as drafts
    • Promote from GitHub: turn a tagged release into a served Relayer release
    • Usage metering: AI quota + storage meters in org settings
  5. Security

    v0.8.0 · Enterprise SSO (beta)

    ssosecurity

    v0.8.0 · Enterprise SSO (beta)

    Bring your identity provider: OIDC SSO with per-org configuration, tested against Microsoft Entra.

    Added

    • SSO provider picker with IdP trust rules and optional email-claim mapping
    • Login with SSO flow alongside existing methods

    Changed

    • Pricing restructured around monthly active devices, flat Team tier, unlimited seats
  6. Security

    v0.7.0 · Accounts, hardening + auth telemetry

    securityaccounts

    v0.7.0 · Accounts, hardening + auth telemetry

    A production-trust release: account controls, evidence trails and a hardening pass audited end to end.

    Added

    • Profile page: auth methods, linked GitHub identity, session management
    • Versioned Terms acceptance log with consent evidence (IP + user agent)
    • Org-visible sign-in event log (sign-ins, failures, sign-outs)

    Changed

    • Redesigned sign-in/up pages, accessibility end to end
    • Production hardening: configurable rate limits, per-account auth limits, error logging
  7. Update

    v0.6.0 · Hosted changelog pages + feeds

    changelogfeeds

    v0.6.0 · Hosted changelog pages + feeds

    Every app gets a shareable public changelog and machine-readable feeds - rollbacks disappear from all of them automatically.

    Added

    • Hosted changelog page per app with human slugs and pagination
    • RSS 2.0 and JSON release feeds per channel, CORS-open, cached
    • "Get updates" popover (RSS / JSON / page link) on the Releases tab
  8. Update

    v0.5.0 · relayerctl, the zero-dependency CLI

    clici

    v0.5.0 · relayerctl, the zero-dependency CLI

    Publish from any CI in one line: npx relayerctl releases create ... with artifacts, channel and starting rollout. One binary, an API key, no SDK.

    Added

    • relayerctl on npm: releases create/list, artifact flags, channel selection
    • Safe retries: duplicate versions answer 409, so re-run CI jobs are harmless
  9. Update

    v0.4.0 · Push-mode fleets

    fleetapi

    v0.4.0 · Push-mode fleets

    For fleets where devices never call home: an orchestrator reports state and plans whole update waves in one call, using the exact same decision logic as /u/.

    Added

    • POST /api/v1/fleet/report - bulk device state, the heartbeat for fleet visibility
    • POST /api/v1/fleet/decisions - up to 500 decisions per call, dry-run mode
    • Device labels: opaque key/values (customerRef, siteRef, ...) stored and echoed
  10. Update

    v0.3.0 · The multi-tenant core

    rolloutsfleetsteams

    v0.3.0 · The multi-tenant core

    The foundation: devices ask GET /u/... "I am X, running Y - what should I do?" and Relayer answers with staged rollouts, channel policy and full fleet visibility.

    Added

    • Update-check endpoint with semver-aware comparison and artifact matching
    • Staged rollouts: sticky per-device cohorts, pause, permanent rollback
    • Channels with min-version policy (below it, updates become required)
    • Fleet dashboard: adoption, 24h check-ins, device pages + detail sheet
    • Tauri dynamic-updater contract + electron-updater feed out of the box
    • Organizations, roles (owner/admin/member), invite links, audit log