esc close · ← → switch
Lyra · Always shipping

Five versions in four months.

A public record of what shipped in Lyra. Curated subset of the engineering log — every entry rewritten in plain language, dated, versioned.

Versions shipped 5
Cadence Monthly
Last shipped April 20, 2026
Range v.2.2.0 → v.2.6.0
Subscribe RSS available
Latest release
v.2.6.0 · April 20, 2026 · PricingFrontend

Self-serve Trial Flow and Side-by-Side Pricing

Launched the self-serve trial flow via Stripe Checkout. "Start Free Trial" buttons on the pricing page now route directly to Stripe with the 14-day trial applied, and `/get-started` has been rebuilt as a plan picker with a 3-step onboarding timeline — Pick your plan, Add your card, Connect your Google Ads.

v.2.6.0 · illus.
+

New Features

2 items

Launched the self-serve trial flow via Stripe Checkout. "Start Free Trial" buttons on the pricing page now route directly to Stripe with the 14-day trial applied, and /get-started has been rebuilt as a plan picker with a 3-step onboarding timeline — Pick your plan, Add your card, Connect your Google Ads.

Redesigned the pricing page with Monthly and Annual plans displayed side-by-side. The billing toggle now emphasises the selected plan rather than hiding the other, so the ~2 months free savings on annual billing is visible at a glance.

~

Improvements

3 items
  1. Side-by-side pricing cards: Monthly ($49) and Annual ($490) shown together with a "Save $98/yr · ~2 months free" badge on the annual card.
  2. Stripe Checkout integration: Pricing and /get-started CTAs route directly to Stripe Payment Links (monthly and annual), with UTM + click IDs forwarded for marketing attribution.
  3. Corrected annual price: Annual billing is $490/year (was previously shown as $470). Monthly usage quotas are unchanged — 500 AI chat messages, 10 campaigns, 100 edits, 30 reports, daily digests.

The rest of 2026, so far.

Click any entry to read the full notes inline. Hash links resolve — share /changelog/#v-2-4-0 and the row stays open on arrival.

April 1, 2026 v.2.5.0

Performance Max AI Intelligence Tab and Batch Distributions Endpoint

Introduced the AI Intelligence tab on the Performance Max page, providing AI-generated insights and optimization recommendations for PMax campaigns. The new batch distributions endpoint consolidates asset performance data into a single API call, reducing page load from 21 HTTP requests down to 4.

AIBackend
+

New Features

2 items

Introduced the AI Intelligence tab on the Performance Max page, providing AI-generated insights and optimization recommendations for PMax campaigns. The new batch distributions endpoint consolidates asset performance data into a single API call, reducing page load from 21 HTTP requests down to 4.

Added lazy-loading for asset groups and server-side pagination for aggregated search terms, significantly improving dashboard responsiveness for accounts with large PMax campaigns.

~

Improvements

3 items
  1. Batch Distributions CTE: Single query replaces multiple individual asset distribution calls, reducing backend load and frontend rendering time.
  2. Search Terms Aggregation: Server-side aggregation with pagination eliminates the need to load all search terms at once.
  3. Asset Group Lazy Loading: Asset groups load on demand rather than upfront, improving initial page render time.
March 1, 2026 v.2.4.0

AI Pattern Learning System with Admin Review and Analytics

Launched the complete AI Pattern Learning System. Lyra now extracts optimization patterns from diary entries and applies them to future recommendations. The system supports multi-strategy pattern retrieval across customer-specific, industry-level, and global contexts.

AIBackend
+

New Features

2 items

Launched the complete AI Pattern Learning System. Lyra now extracts optimization patterns from diary entries and applies them to future recommendations. The system supports multi-strategy pattern retrieval across customer-specific, industry-level, and global contexts.

Added the admin review interface at /admin/ai-patterns for approving extracted patterns and the analytics dashboard at /admin/pattern-analytics for tracking pattern usage and effectiveness.

~

Improvements

3 items
  1. Pattern Extraction Pipeline: Automated extraction from diary entries with pending/approved workflow.
  2. Vector Search Integration: Patterns stored with embeddings for semantic similarity matching using pgvector.
  3. Usage Tracking: Pattern application counts and feedback ratings tracked for continuous improvement.
February 1, 2026 v.2.3.0

Campaign Health Analyzer v2.1.0 with Quality Score Diagnostics

Upgraded the Campaign Health Analyzer to v2.1.0 with integrated Quality Score diagnostics. The analyzer now breaks down QS components including expected CTR, ad relevance, and landing page experience at the keyword level.

AIFrontend
+

New Features

2 items

Upgraded the Campaign Health Analyzer to v2.1.0 with integrated Quality Score diagnostics. The analyzer now breaks down QS components including expected CTR, ad relevance, and landing page experience at the keyword level.

Added performance labels for campaigns and ad groups, providing at-a-glance status indicators with severity-based prioritization.

~

Improvements

3 items
  1. QS Component Breakdown: Individual scoring for each Quality Score factor with specific improvement recommendations.
  2. Performance Labels: Color-coded severity labels applied to campaigns, ad groups, and keywords.
  3. Prioritized Recommendations: Optimization suggestions ranked by potential impact and implementation effort.
January 1, 2026 v.2.2.0

AI Bid Optimization Engine v2.0.0

Released the AI Bid Optimization Engine v2.0.0 for Manual CPC campaigns. The engine provides keyword-level bid adjustment recommendations with real-time performance prediction and configurable safety controls.

AIBackend
+

New Features

2 items

Released the AI Bid Optimization Engine v2.0.0 for Manual CPC campaigns. The engine provides keyword-level bid adjustment recommendations with real-time performance prediction and configurable safety controls.

Added bid simulation capabilities that show projected performance changes before applying adjustments, reducing the risk of over-bidding or under-bidding on high-value keywords.

~

Improvements

3 items
  1. Performance Prediction: Projected click, cost, and conversion estimates for proposed bid changes.
  2. Safety Controls: Configurable bid caps, daily budget limits, and rollback thresholds to prevent runaway spend.
  3. Batch Operations: Apply bid adjustments across multiple keywords and campaigns in a single action.
RSS · feed

Read it in your reader.

Same data, machine-readable. Auto-generated from the source YAML — every entry on this page is in the feed.

Subscribe via RSS /changelog/rss.xml
Email · monthly

Or get a single email per release.

Monthly emails when we ship — no marketing, no roadmap teasers, no "you might also like". Just the changelog. Unsubscribe is one click in every email.

14-day free trial · No credit card · Cancel in one click

Run the audit.
Keep the findings.

Connect your account and let Lyra run its 18 tools for 14 days. If the projected waste recovery isn't worth at least 10× the $49, don't pay. You keep every insight either way.

Start 14-day free trial

Read-only connect · write access opt-in per tool · SOC 2 in progress · GDPR + CCPA compliant