Changelog

All notable changes to AgentLeverage are documented here.

v1.8.1October 2026Quote Check — verify minutes against the recording
  • NewCheck quotes, decisions, and action items in finished minutes against the labeled Speaker Separation transcript. Each item comes back verified, paraphrase-supported, or unsupported / misattributed. No re-upload. 2 credits per run
  • NewQuote Check is on MCP as quote_check (tools:execute). Pass sourceJobId for a completed meeting_summarizer job that chained from speaker_separation, or a pasted summary plus transcriptJobId. The minutes and transcript jobs are not changed. Agents still cannot delete jobs, files, or tags over MCP
v1.8.0October 2026Recording to minutes, in the browser and over MCP
  • NewCreate minutes from a finished Speaker Separation job — summary, decisions, and action items with owners and deadlines when those were said. No re-upload. 2 credits per run, same as pasting notes into Meeting Summarizer
  • NewMeeting Summarizer is on MCP as meeting_summarizer (tools:execute). Pass meetingNotes or sourceJobId for a completed speaker_separation job in the same organization. The source job is not changed. Agents still cannot delete jobs, files, or tags over MCP
v1.7.0September 2026Agents can find org files over MCP
  • NewMCP agents can list organization files, mint a download URL, rename speaker labels, and read credit history — so a request like “run speaker separation on the standup recording” no longer needs the storage path pasted into the prompt
v1.6.0August 2026Audio Redaction, Background Remover & speaker-separation takeaways
  • NewAudio Redaction — strip names, numbers, and other PII from a recording; get a redacted transcript plus bleeped audio (3 credits/run)
  • NewBackground Remover — remove image backgrounds with AI, upload a file or pass a public HTTPS URL (1 credit/image)
  • NewSpeaker Separation results now include copy-to-clipboard and txt/srt downloads alongside the existing transcript view
  • NewSpeaker Separation emails you when a job finishes, so you don't have to keep the tab open
  • ImprovedAfter your first tool run, we suggest a next tool or agent connection instead of leaving you on a blank dashboard
  • ImprovedOrg-scoped static API tokens: agl_ tokens now act as org credentials, so any org member's job runs are visible to the rest of the org instead of siloed to the token creator
  • ImprovedMCP server code for Claude custom-connector OAuth is on main; going live still needs Clerk Dashboard Dynamic Client Registration, so token-based MCP access remains the supported path for now
v1.5.0April 2026Monetization, analytics & accessibility pass
  • ImprovedStarter→Pro upgrade nudges added to every high-traffic surface: dashboard, billing settings, empty states, post-job prompts, credit banners, and history pages
  • ImprovedAnnual billing upsell card in billing settings for Pro monthly subscribers — shows computed savings and switch-to-annual CTA
  • NewPaymentIssueAlert surfaces subscription payment failures (past_due, unpaid, paused) on dashboard and billing settings to reduce involuntary churn
  • NewWin-back nudge in billing settings for canceled or expired subscriptions
  • NewUsage overage nudge in UsageSummaryCards: Starter/Free users with overage charges see a quantified Pro upgrade prompt
  • ImprovedTrust signals (30-day money-back, cancel anytime, secure payment) added to the choose-plan forced conversion gate
  • ImprovedExpanded analytics instrumentation across conversion, auth, navigation, dashboard, and tool surfaces
  • ImprovedWCAG 2.1 accessibility fixes: aria-labels for icon-only buttons, aria-current for active items, role/label for tables, and accessible password toggle in auth forms
  • ImprovedError boundaries added to all major app routes; loading skeletons added to settings, admin, tool, and dashboard routes
  • ImprovedToolHistoryPage now uses server-side pagination (20 per page) instead of bulk 100-job fetch
  • FixedFixed hardcoded tool-slug→detail URL maps in ToolRecentRuns and ActiveJobsWidget — all tools now link to the correct job detail page
  • FixedFixed money-back guarantee inconsistency across trust surfaces (standardized to 30-day everywhere)
  • FixedFixed StickyCta and BlogStickyCta showing free-trial copy to Starter plan users who are already paying
  • ImprovedPricing FAQ expanded from 6 to 11 entries — covers Starter vs Pro, plan switching, team pricing, file formats, and guarantee
  • ImprovedSEO: added JSON-LD structured data to /for, /vs, /roi-calculator, and /use-cases pages
v1.4.0March 2026AI-powered dashboard & UX overhaul
  • NewAdded personalized dashboard with activity heatmap, streak widget, daily goals, and credit forecasting
  • NewAdded job comparison, pinned jobs, job notes, and job tagging to the jobs history page
  • NewAdded ToolPreviewDrawer, ToolTipsBanner, ToolUsageGuide, ToolSampleOutput, and ToolBenchmarkWidget
  • NewAdded keyboard navigation shortcuts (g+h/t/j/s) and keyboard shortcuts help dialog (?)
  • NewAdded searchable, filterable, and sortable ToolsGrid with favorites and tool collections
  • NewAdded milestone notifications, job completion toasts, and credit alert settings
  • ImprovedRedesigned job detail page with output viewer, run-again action, and auto-polling
  • ImprovedBulk select and delete jobs from the jobs history page
v1.3.0March 2026SEO & marketing pages
  • NewAdded public marketing landing pages for each AI tool with hero, features, use-cases, and JSON-LD structured data
  • NewAdded dynamic OG image generation endpoint for social sharing
  • NewAdded BreadcrumbList, FAQPage, and Article JSON-LD structured data throughout the site
  • ImprovedUpdated robots.txt and sitemap.xml to include all public tool pages
  • ImprovedAdded social proof bar and three new blog posts on AI tool use cases
v1.2.0February 2026Test coverage & reliability
  • ImprovedExpanded automated test coverage across web, API, database, storage, and mail packages
  • ImprovedDockerless test runner ergonomics — all tests skip or skip-integrate gracefully when Docker is unavailable
  • ImprovedRemoved host psql dependency from local-eval, setup, and seed flows
  • ImprovedSignificant speedups in test suites via parallelization and shared fixtures
v1.1.0January 2026Local development & onboarding
  • NewAdded pnpm local-eval:smoke command for single-command local evaluation
  • ImprovedTightened README and CONTRIBUTING docs — local evaluation path now reflects real prerequisites
  • ImprovedAdded pnpm worktree:create / resume / list / remove for parallel feature development
v1.0.0December 2025Initial launch
  • NewMeeting Summarizer — AI-powered meeting notes and action item extraction
  • NewInvoice Processor — extract line items, totals, and vendor details from invoice PDFs
  • NewSpeaker Separation — diarize audio into per-speaker segments
  • NewNews Analyzer — curate and analyze news articles on any topic
  • NewCredit-based pricing with free tier, Starter plan, and Pro plan
Changelog — What's New | AgentLeverage