CHANGELOG

What's new in yeet.seo

Every feature, improvement, and fix — shipped fast.

v0.40.0Latest

Skills: Agent-Friendly Composable Workflows

  • AddedSkills system -- high-level composable workflows that bundle Yeet's 60+ tools into single operations. External agents (OpenClaw, Hermes, Claude Code) call one skill instead of orchestrating dozens of tools.
  • Addeddaily_digest skill -- a prioritized daily summary of fixes to apply, content opportunities, verified wins, rank movers, alerts, and indexing coverage. Pre-computed at 5:30am UTC so agents get instant results.
  • Addednew_content skill -- end-to-end content creation: auto-discovers highest-ROI topics from GSC + competitor gaps, builds a strategist brief with SERP analysis and AEO data, optionally generates a full draft.
  • Addedsite_pulse skill -- quick health check showing indexing coverage, rank distribution, AI citation counts, and active optimization cards for any connected property.
  • AddedMulti-format output -- every skill renders as rich markdown (CLI/IDE), compact chat text (Telegram/Discord), or structured JSON (dashboard). Beautiful progress bars for indexing coverage across all surfaces.
  • Added3 new MCP tools: list_skills (discover available workflows), run_skill (execute with format selection), get_skill_result (fetch cached results). Plus a latest-digest MCP resource for autonomous agents.
  • AddedREST API: GET/POST /api/v1/skills for listing, running, and fetching cached skill results with format parameter (json, markdown, chat).
  • Addedskill_runs table stores results with pre-rendered markdown and chat formats for async retrieval by agents polling later.
v0.39.0

Headless-First MCP + Activation

  • AddedFree tier now includes 2 SEO agent runs on signup. The first run kicks off automatically when you finish OAuth -- first findings appear in ~10 minutes.
  • AddedSignup auto-trigger: connecting Google Search Console immediately queues a full pipeline pass (sitemap sync + SEO agent + insights). No more waiting until 4am UTC for the first run.
  • AddedHeadless OAuth device flow: POST /api/auth/device/{start,authorize,poll} lets CLIs and MCP clients bootstrap an API key without copy-pasting from a browser settings page.
  • AddedWebhooks: subscribe to card.created, card.applied, agent_run.completed, rank.changed, citation.detected, and more. HMAC-signed deliveries via X-YeetSeo-Signature.
  • AddedSSE event stream at GET /api/v1/events/stream for real-time push without webhooks.
  • AddedIdempotency-Key header support on every write endpoint -- safe to retry from headless agents.
  • AddedMCP server gains 18 new tools (properties connect/disconnect, api keys CRUD, team mgmt, scheduler config, pipeline control, agent trigger, card lifecycle, webhooks, credits), 6 new resources, and 6 prompt templates (full_site_audit, fix_priorities, content_brief, competitor_gap_analysis, ai_citation_audit, headless_bootstrap).
  • AddedStreamable HTTP transport for the MCP server (set YEET_SEO_MCP_TRANSPORT=http) so it can run as a hosted service alongside stdio.
  • ImprovedScheduler manual_trigger now fans out per-user via a queue table and dispatches via triggerAgentJob -- no more global clobbering when multiple users trigger simultaneously.
v0.38.0

Autonomous Agent Loop

  • AddedAutonomous agent loop -- Yeet now continuously evaluates workspace freshness and queues the next highest-priority task (sitemap refresh, SEO agent, investigator, ideator) without manual intervention.
  • AddedAutonomous dashboard -- the old "Submit pages" panel is replaced with a live status view showing Yeet's current mode, reasoning, next planned actions, active work, and pending approvals.
  • AddedNew API endpoints: GET /api/v1/autonomous/state returns the full autonomous evaluation, POST /api/v1/autonomous/run triggers a cycle manually.
  • AddedMCP tools: autonomous_state and autonomous_run let AI agents query and trigger the autonomous loop from Cursor, Claude Code, or any MCP client.
v0.37.0

Sentry Error Tracking & Retry Resilience

  • AddedSentry integration -- scheduler and Next.js both report errors to Sentry in production. Every cron job and unhandled rejection is tagged and tracked.
  • ImprovedRetry with exponential backoff for Perplexity, Gemini, and other external API calls. Transient failures no longer break agent runs.
  • ImprovedGemini client retries on 429/5xx with jittered backoff and respects Retry-After headers.
v0.36.0

Agent-First API & MCP Expansion

  • AddedWorkspace overview API -- GET /api/v1/workspace/overview returns property counts, card breakdowns, recent agent runs, scheduler health, and sitemap snapshots in one call.
  • AddedAgent run API -- POST /api/v1/agent/run triggers full agent runs headlessly. GET /api/v1/agent/runs returns run history.
  • AddedScheduler control API -- GET /api/v1/system/scheduler returns scheduler health and config.
  • Added12 new MCP tools for workspace overview, agent runs, scheduler health, sitemap snapshots, and feed access.
v0.35.0

Agentic Insights Pipeline

  • AddedInvestigator agent -- a Gemini 2.5 Pro tool-use agent that scrapes your pages, compares against SERP competitors, and promotes the strongest findings into actionable fix cards.
  • AddedVerifier agent -- a state-machine that re-scrapes pages after changes are applied, confirms whether a fix actually landed, and moves cards through in_progress, under_review, verified, or regressed states.
  • AddedContent ideator -- discovers high-leverage content gaps and AEO/GEO opportunities, creating "create" cards alongside the investigator's "fix" cards.
  • AddedFirecrawl scraper -- caches page snapshots, diffs before/after changes, and supports full-site crawl for sitemap discovery.
  • AddedUnified activity feed -- GET /api/v1/feed returns a chronological feed of agent runs, card state changes, and submissions with lifecycle badges.
  • ImprovedOptimization engine now writes to an intermediate candidates table. Only the investigator promotes candidates to user-facing cards, reducing noise.
  • ImprovedUpgraded to Gemini 2.5 Pro with native tool-use loop and thought signatures for more accurate, citation-backed recommendations.
  • FixedCard UI deduplication -- resolved insights no longer reappear after a property refresh.
v0.34.0

Team Workspaces

  • AddedTeam invites -- Pro users can invite up to 3 team members to share their workspace, insights, optimization cards, and credit pool.
  • AddedShared credit pool -- all team members draw from the owner's credit balance. Credit transactions track who spent each credit.
  • AddedTeam management in Settings -- invite by email, view members, copy invite links, remove members, or leave a team.
  • AddedAutomatic team join on login -- invited users are added to the team when they sign in with Google, no extra confirmation needed.
v0.33.0

Smart Card Resolution & In-Progress Tracking

  • AddedIn-Progress card state — when a fix is detected in your source code (via GitHub) but not yet reflected in analytics, the card shows an "In Progress" badge instead of Applied/Dismiss buttons.
  • AddedGitHub code scanning for auto-resolve — searches recent commits and source code for evidence that a recommended fix has been implemented.
  • ImprovedAuto-resolve now promotes in-progress cards to resolved once analytics confirm the change is live.
  • ImprovedCards API supports multi-status filtering so the dashboard shows both active and in-progress cards together.
v0.32.0

Vibe Max Overhaul, MCP Vibe Fix & Agent Workflow

  • AddedVibe Max MCP tool — AI agents can call vibe_max to get all SEO fixes combined into a single mega-prompt. Works from Cursor, Claude Code, or any MCP client.
  • Addedget_vibe_fix MCP tool — get or generate the vibe fix prompt for a specific insight. Returns instantly if already generated, otherwise generates it (1 credit) and returns the result.
  • AddedVibe Max REST API — GET /api/v1/vibe-max returns the assembled mega-prompt via API key auth. POST /api/v1/vibe-fix generates individual insight fixes headlessly.
  • ImprovedVibe Max now batch-generates vibe fixes for all insights before assembling. Each insight card shows a spinner during generation and flips to a checkmark when done.
  • ImprovedVibe Max preview dropdown shows full credit breakdown (ready fixes free, insights to generate at 1 credit each) before the user confirms.
  • ImprovedAfter Vibe Max completes, a notification bar appears with Copy and Download .md buttons instead of auto-copying. Large outputs (>50k chars) hide the copy button and show download only.
  • Improved"Copy recent" option in the Vibe Max dropdown lets users re-access the last generated output without regenerating.
  • ImprovedAll insight card bolt icons update to copy icons after vibe fix generation, whether via individual click, Vibe Max batch, or MCP.
  • FixedVibe Max no longer outputs "undefined" for insights without vibe fix prompts. Every item in the output is a real, actionable fix.
v0.31.0

Documentation Rewrite & Full Feature Catalog

  • ImprovedComplete rewrite of all user documentation: Overview, Agent & Chat, API, and Integrations pages now cover content intelligence, Vibe Max, all 37 MCP tools, and the agent workflow guide.
  • AddedAgent Workflow Guide in docs — step-by-step instructions for AI agents to discover issues, get fix prompts, apply changes, and track progress using MCP tools.
  • AddedAPI docs now include all 45+ REST endpoints organized by category with curl examples for content intelligence, vibe fix, and vibe max.
v0.30.0

Content Intelligence — SERP Analysis, Topic Scoring & AI Content Generation

  • AddedSERP Analyzer — analyze the top 10 Google results for any keyword. Extracts headings, topics, word counts, and questions from competitors to build a data-driven content strategy.
  • AddedTopic Scoring Engine — score any page 0-100 on topic coverage compared to SERP competitors. Shows covered terms, missing terms, and specific suggestions to improve coverage.
  • AddedAI Content Brief Generation — generates SEO+AEO content briefs from SERP analysis. Includes target word count, suggested title, headings, topics to cover, questions to answer, and AEO schema recommendations.
  • AddedAI Content Generation — writes full SEO+AEO optimized articles from a keyword. Iteratively scores against the topic model and rewrites until the quality threshold is met. Outputs publish-ready markdown.
  • AddedExisting Content Optimizer — analyzes any URL against its keyword's SERP competitors and generates specific improvement suggestions: paragraph additions, heading rewrites, FAQ sections, and projected score after changes.
  • Added5 new MCP tools: analyze_serp, score_content, generate_brief, generate_content, optimize_content. Use from Cursor, Claude Code, or any MCP client.
  • Added5 new REST API endpoints: /api/v1/content/{analyze-serp, score, brief, generate, optimize}. Full headless content intelligence.
  • AddedContent Gap Watcher — new pipeline watcher runs daily, identifies page-2 keywords without dedicated pages, and auto-generates optimization cards with content creation opportunities.
  • AddedContent drafts table — generated content is saved as drafts with topic scores, covered/missing terms, and brief metadata for future editing and publishing.
v0.29.0

GitHub Integration, Vibe Max & Auto-Discovery

  • AddedGitHub read-only integration — connect your GitHub account to cross-reference source code in optimization prompts. Select a repo from Settings, and Vibe Fix prompts include file paths and framework context.
  • AddedAuto-discover keywords and competitors — when you connect a property, keyword tracking and competitor lists are automatically populated from your existing GSC data. No manual input needed.
  • AddedVibe Max button — one-click mega-prompt that combines all optimization cards and insights into a single, priority-sorted IDE-ready prompt for Cursor or Claude Code.
  • AddedAuto-resolve optimization cards — cards are automatically marked as resolved when a data refresh confirms the underlying issue has been fixed on the live site.
  • AddedMCP: GSC search analytics — new MCP tools expose search queries, page performance, impressions, clicks, and CTR data for external AI clients.
  • AddedMCP: DataForSEO page data — keyword and backlink data for specific pages now available via MCP tools.
  • AddedRate limiting for URL submissions — per-user daily submission caps with witty feedback messages to prevent quota exhaustion.
  • ImprovedVibe Max dropdown overhauled — shows a breakdown of ready fixes (prompt already generated) vs new fixes (need generation), so you know exactly what you're getting before clicking.
  • FixedInsights generation no longer gets stuck after a 5-minute timeout — the generator now properly aborts and retries on the next cycle.
  • FixedHelpful message shown when insights need Vibe Fix generated before Vibe Max can include them, instead of silently excluding them.
v0.27.1

Data-Rich Vibe Fix, Cairo CDP & Bug Fixes

  • AddedData snapshot caching — insights generation now captures a full snapshot of GSC analytics, GA4 traffic, backlink profile, rank tracking, LLM mentions, and competitor data alongside insight cards. Vibe Fix prompts reference real numbers (impressions, CTR, positions, search volume) instead of generic advice.
  • AddedCairo CDP event tracking — user signups, checkout starts/completions, insights generated, Vibe Fix generated, and agent runs are now tracked via a self-hosted CDP with Discord bridge notifications for key events.
  • AddedNew playbook article: "Programmatic SEO vs Traditional SEO" — a MOFU post by Nova Ellis covering when to use each approach.
  • ImprovedCopy-prompt and Vibe Fix outputs now inject real user data (affected URLs, GSC metrics, rank positions, backlink counts) making the prompts immediately actionable in Cursor or Claude Code.
  • FixedInsight filtering moved to server-side — removes client-side filtering inconsistencies and ensures insights match connected properties correctly.
  • FixedError agent runs are now excluded from insights generation context, preventing stale or failed data from influencing recommendations.
  • FixedSmart hostname-based insight filtering — sc-domain URL normalization now handles www. variants, fixing invisible insights for some GSC properties.
  • FixedCloud Build deploy now uses git reset --hard before pull to prevent merge conflicts on the VM.
v0.27.0

Autonomous Optimization Pipeline & Unified Feed

  • AddedOptimization Cards system — discrete, queryable cards replace the monolithic insights cache. Each card has a category (FIX or CREATE), priority, confidence score, evidence, and an IDE-ready prompt you can paste directly into Cursor or Claude Code.
  • AddedContinuous Pipeline — 8 source-specific watchers (GSC, GA4, backlinks, rank tracking, competitors, LLM mentions, citations, prompt rank) run on adaptive intervals instead of once-daily cron. New data automatically generates optimization cards.
  • AddedReal-time SSE feed — new cards stream to the dashboard instantly via Server-Sent Events. No more refreshing to see new recommendations.
  • Added5 new MCP tools: get_optimization_cards, get_priority_actions, apply_card, dismiss_card, get_card_prompt. Plus 2 new resources: yeet-seo://optimization-feed and yeet-seo://content-gaps.
  • AddedCards REST API — GET /api/v1/cards (list/filter), POST /api/v1/cards/:id/apply, POST /api/v1/cards/:id/dismiss, GET /api/v1/cards/:id/prompt.
  • Added7 deterministic card generators: meta tag fixes, rank recovery, backlink loss alerts, content gap opportunities, technical SEO issues, keyword cannibalization, and page-2 push opportunities.
  • AddedSEO agent now emits optimization cards alongside hypotheses — every finding becomes an actionable card in the unified feed.
  • ImprovedDashboard redesigned with a single unified "What to do next" feed — optimization cards and AI insights merged into one priority-sorted stream. No more separate sections.
  • ImprovedInsight cards now show typed badges (FIX, WARNING, WIN, INFO) with color-coded left borders instead of plain dots. Visual style matches optimization cards for a consistent feed.
  • ImprovedRemoved SEO Tools grid from dashboard — the product is the agent, not a collection of manual tools. Keywords, backlinks, rank tracking, and competitor pages remain accessible from the nav.
  • ImprovedPipeline respects plan gating — free users get cards from GSC, GA4, and citation watchers; DataForSEO-powered watchers only run for paid plans.
  • ImprovedFingerprint-based card deduplication — same issue detected again supersedes the stale card instead of creating duplicates. 6-hour cooldown prevents noise.
v0.25.0

LLM Mentions API & Enhanced AI Insights

  • AddedDataForSEO LLM Mentions integration — monitor brand visibility across ChatGPT and Google AI Overviews using structured data from 188M+ indexed LLM prompts. Replaces basic Perplexity/DuckDuckGo fallback with real AI visibility intelligence.
  • Added4 new LLM Mentions API endpoints: /api/v1/llm-mentions/search, /metrics, /top-domains, /top-pages.
  • Added3 new MCP tools: get_llm_mentions, get_llm_mention_metrics, get_llm_top_domains — AI agents can now query LLM visibility data directly.
  • ImprovedAI Insights now generate 8-15 cards (up from 5-7) covering all data sources. Cards are ordered by highest priority and business impact — anomalies first, then growth opportunities, then maintenance.
  • ImprovedInsights now incorporate enriched keyword data (volume, KD, CPC cross-referenced with rank positions), LLM mention volume by platform, DataForSEO usage tracking, and page-2 keyword opportunities.
  • ImprovedCitation monitoring worker now auto-selects the best data source: DataForSEO LLM Mentions API when configured, Perplexity API as secondary, DuckDuckGo HTML as fallback.
v0.24.0

Autonomous SEO Agent with Hypothesis System

  • AddedAutonomous SEO Intelligence Agent — Gemini-powered agent that proactively analyzes GSC queries, backlinks, competitors, and rank data to generate hypotheses and take action without user input.
  • AddedThree-tier hypothesis system: Tier 1 (auto-act silently on validated data), Tier 2 (act and notify with hypothesis cards), Tier 3 (propose and wait for user approval before spending).
  • AddedHypothesis categories: keyword opportunities, content gaps, cannibalization, competitor moves, rank drops, link opportunities, trending queries, backlink loss, adjacent topics, new verticals, deep audits.
  • AddedHypothesisCard UI component — color-coded by tier (green=auto-acted, yellow=hypothesis, blue=needs approval) with confidence scores, evidence, and approve/dismiss buttons.
  • AddedHypotheses API at /api/v1/hypotheses — list, filter, approve, and dismiss hypotheses. 3 new MCP tools for hypothesis management.
  • AddedAgent bootstraps automatically when a user connects a new GSC property — seeds keywords, discovers competitors, snapshots backlinks, and generates initial hypotheses within minutes.
  • ImprovedAll SEO pages (Keywords, Backlinks, Rank Tracking, Competitors) rewritten as auto-populated dashboards showing agent findings and hypothesis cards instead of empty input forms.
  • ImprovedDaily 4am UTC cron runs the SEO agent for all non-free users with connected properties.
v0.23.0

DataForSEO Integration: Keywords, Backlinks, Rank Tracking, Competitors

  • AddedDataForSEO API integration — keyword research (volume, KD, CPC, intent, suggestions), backlink analysis (profile, referring domains, anchors, history, competitors), SERP rank tracking (daily positions), and competitor analysis (organic competitors, keyword gaps, domain overview, shared backlinks).
  • Added18 new REST API endpoints under /api/v1/keywords, /backlinks, /rank-tracking, and /competitors.
  • Added15 new MCP tools for keyword research, backlink analysis, rank tracking, and competitor analysis — available in Cursor, Claude Desktop, and any MCP client.
  • Added14 new chat agent tools — ask about keywords, backlinks, competitors, and rankings in natural language via the chat interface.
  • AddedDaily background workers: backlink snapshots at 3am UTC, SERP rank checks at 3am UTC. DataForSEO usage tracking with per-endpoint cost logging.
  • Added4 new frontend pages: Keywords, Backlinks, Rank Tracking, Competitors — accessible from the dashboard SEO Tools grid.
  • AddedPlan gating: keyword research on Basic plan, backlinks/rank tracking/competitors on Pro plan.
v0.22.0

AI-Native SEO: Discoverability, Citations, Prompt Rank & Auto-Fix

  • AddedDiscoverability Score — unified 0-100 score per page combining indexing status, AEO readiness, schema presence, and search performance. Available via API, MCP, and agent tools.
  • AddedAI Citation Monitoring — daily background scans detect when ChatGPT, Perplexity, Claude, and Gemini mention your site, even without click-through traffic. View citations via API or MCP.
  • AddedPrompt Rank Tracking — track where your product appears when specific prompts are asked to AI models. Add prompts to monitor, see rank position and citation status over time.
  • AddedAuto-Fix Pipeline — generate structured, ready-to-apply SEO fixes (JSON-LD schema, meta tags, llms.txt, internal link suggestions) for any URL. Available via API and MCP.
  • AddedDeploy Notify endpoint — unified POST /api/v1/deploy-notify pipeline that inspects, submits, AEO-checks, and scores discoverability for URLs on every deploy.
  • Added5 new MCP tools: get_discoverability_score, get_ai_citations, track_prompt, get_prompt_ranks, generate_auto_fix.
  • ImprovedGitHub Action now supports discoverability-check input — calls deploy-notify and reports discoverability scores in the GitHub step summary.
  • ImprovedNext.js webhook plugin updated to call deploy-notify and return per-URL discoverability results, with Vercel deployment payload support.
  • ImprovedAI Insights now incorporate citation volume, prompt rank visibility, and AI discovery signals when generating recommendations.
v0.21.0

AI Insights via MCP + Dashboard UX Overhaul

  • AddedNew get_insights MCP tool — AI agents in Claude Code, Cursor, Windsurf, or any MCP client can now pull prioritized SEO recommendations and act on them autonomously.
  • AddedNew /api/v1/insights endpoint for programmatic access to AI-generated SEO insights via API key.
  • AddedScheduler health system — heartbeat monitoring detects when the scheduler process is offline and surfaces it in the dashboard.
  • AddedSubmission history page at /activity with full agent run logs and error highlighting.
  • AddedMCP client compatibility logos (Claude, Cursor, Windsurf, Cline, Duet) shown in Settings API Keys section.
  • ImprovedDashboard: replaced noisy agent activity log with a compact Page Submissions health card showing status, last run stats, and a Submit Pages button with real-time status polling.
  • ImprovedRenamed 'Force refresh data' to 'Refresh insights' and moved it next to the AI Insights section where it belongs.
  • ImprovedAPI key creation now supports custom names (e.g. Local, CI). MCP config snippet uses a clean placeholder instead of masked key prefixes.
  • FixedAI insights were invisible for sites using sc-domain: GSC properties — the connected URL filter now normalizes domain properties to match insight siteUrls.
  • FixedRun Indexer button now checks scheduler health before queuing and shows a clear error when the scheduler is offline.
v0.18.0

Pro Plan Gating for API Keys & MCP

  • AddedAPI key creation and deletion now gated behind the Pro plan — free and basic users see an upgrade prompt with a direct link to /pricing.
  • AddedSettings page has a new API Keys & MCP section — Pro users get full key management UI and a ready-to-paste MCP config snippet; non-Pro users see an upgrade CTA.
  • ImprovedPro plan feature list on the Pricing page now explicitly calls out 'API access + API key management' and 'MCP server' as Pro-only features.
v0.17.0

Bug Fixes: DB Auth & Agent Tools

  • FixedagentTools.js getOAuth2Client now handles session tokens as either a JSON string or an already-parsed object — fixes a crash when pg returns jsonb columns as objects.
v0.16.0

Security: SQL Injection Fix in Indexing Trends

  • FixedParameterized the SQL INTERVAL in getIndexingTrends — the 'days' parameter was previously interpolated directly into the query string, allowing SQL injection via the get_indexing_trends agent tool. Now parsed as integer, clamped to 1–365, and passed as a parameter.
v0.15.0

SEO & Sitemap

  • AddedDynamic XML sitemap at /sitemap.xml covering all public pages with proper lastmod, changefreq, and priority.
  • Addedrobots.txt — allows crawling of public pages, blocks authenticated routes (/home, /settings, /chat) and API endpoints.
  • AddedGlobal _document.js with lang="en", Open Graph and Twitter Card meta tags, theme-color, and favicon across all pages.
  • AddedStructured data on the landing page — WebSite and SoftwareApplication JSON-LD schemas for rich search results.
  • ImprovedCanonical URLs and OG/Twitter meta added to all public pages (landing, pricing, changelog).
  • ImprovedAll authenticated pages (dashboard, settings, chat, login, signup) now have noindex/nofollow to prevent accidental indexing.
  • ImprovedPricing page meta description rewritten with specific plan details and pricing for better SERP snippets.
v0.14.0

Dashboard Status Panel & Plan-Aware Pricing

  • AddedData source status badges on the dashboard — GA4, Search Console, ChatGPT, Perplexity, Claude, and Bing shown with brand logos and a pulsing status indicator.
  • ImprovedPricing page now reflects your current plan — your active subscription is highlighted with "Current plan", and downgrade buttons are disabled.
  • ImprovedAI insights are now filtered by connected properties — hiding a property in Settings also hides its insights from the dashboard.
  • ImprovedNext refresh countdown now shows relative time (e.g. "14h 23m") instead of a full date string.
  • FixedFixed "Cannot access fetchInsights before initialization" error on the home page caused by a temporal dead zone issue.
v0.13.0

Durable Background Jobs

  • AddedGoogle Cloud Tasks integration — insights generation and Vibe Fix jobs are now enqueued as durable tasks with automatic retries (3 attempts, 10s backoff).
  • ImprovedBackground jobs survive process restarts — if the server crashes mid-generation, Cloud Tasks retries the job automatically instead of losing work.
  • ImprovedWorker endpoints with shared-secret auth — dedicated /api/workers routes for Cloud Tasks to call, secured by a secret header.
  • ImprovedZero-setup local development — background jobs run in-process when Cloud Tasks is not configured, so local dev needs no GCP dependencies.
v0.12.0

On-Demand Vibe Fix & Credit Safety

  • ImprovedVibe Fix is now generated on-demand when the copy button is clicked, not pre-generated with insights. Saves credits by only generating what users actually need.
  • ImprovedCredit refund on failure — if Vibe Fix generation fails, the consumed credit is automatically refunded with a transaction record.
  • ImprovedGenerating state tracked per-insight — dashboard shows which specific insight card is generating its Vibe Fix, with real-time polling.
v0.11.0

Plans, Pricing & Mobile

  • AddedPlans & subscriptions — Free, Basic, and Pro plans with Stripe Checkout. Basic ($9/mo) and Pro ($24/mo) with monthly or annual billing; webhooks grant included credits and sync plan status.
  • AddedCredits slider — top up 10–500 credits with a slider and tiered pricing ($0.50/$0.40/$0.30 per credit). One-time payment; no fixed packs.
  • AddedAI insights filter — filter insights by property. Dropdown shows “All properties” or each visible property (from Settings); insights can be tagged by property.
  • ImprovedPricing page — How it works moved to top with 3 columns (icon + text). Full-width credits slider. Stripe setup documented in docs/STRIPE_SETUP.md.
  • ImprovedMobile-friendly nav — hamburger menu on small screens with slide-down links; desktop layout unchanged. Escape and route change close the menu.
v0.10.0

Stripe Credits & Insights Refresh

  • AddedStripe integration — purchase credits to generate Vibe Fix prompts. Checkout flow, webhooks, and transaction history.
  • AddedCredits system — each Vibe Fix generation consumes credits. Credit balance shown on dashboard with quick link to buy more.
  • AddedManual insights refresh — force-refresh AI insights from the dashboard. Data updates on demand instead of waiting for cache expiry.
  • ImprovedScheduled refresh display — dashboard now shows when the next automatic refresh will run (daily at 2am UTC).
  • ImprovedInsights cache status tracking — background generation state is now tracked for more reliable loading and polling.
v0.9.0

Vibe Fix & Background Insights

  • AddedVibe Fix — one-click copy of AI-agent-ready prompts from every insight card. Paste into Cursor or Claude Code to fix SEO issues instantly.
  • AddedBackground insight generation — insights are now generated asynchronously and cached. Page loads are instant; AI generation runs in the background.
  • ImprovedUpgraded insight generation to Claude Sonnet 4.6 with two-phase architecture: fast insight cards first, detailed Vibe Fix prompts second.
  • ImprovedAdded shimmer skeleton loader for insights while AI is generating.
  • ImprovedSEO tips in the loading screen now display for 7 seconds each for better readability.
v0.8.0

GA4 Integration & Search Analytics

  • AddedGoogle Analytics 4 integration — connect GA4 properties to see page traffic, engagement, and AI referral sources (ChatGPT, Perplexity, Claude).
  • AddedGSC Search Analytics — see clicks, impressions, and CTR per page directly in the dashboard. Identify high-impression, low-CTR pages.
  • AddedAI Insights dashboard — Claude-powered analysis of your indexing data with actionable recommendations.
  • ImprovedHome page redesigned with dynamic SEO tips during loading and improved data layout.
  • FixedGA4 integration connection flow, logout handling, and settings status display.
v0.7.0

AEO Analysis & Structured Data

  • AddedDeep AEO (Answer Engine Optimization) analysis — AI-readability scoring with A-F grades, breakdown by title, meta, headings, content depth, and structured data.
  • AddedAutomatic structured data generation — FAQPage, HowTo, and Article JSON-LD suggestions based on page content.
  • Addedllms.txt generation — create machine-readable site descriptions for AI crawlers.
  • AddedBatch AEO scoring across multiple URLs.
v0.6.0

Developer Integrations

  • AddedGitHub Action — auto-submit new pages on every deploy, with optional AEO quality gates.
  • AddedNext.js plugin (@yeet-seo/nextjs) — drop-in JSON-LD schema components, automatic llms.txt serving, and post-deploy indexing hooks.
v0.5.0

CLI, MCP Server & REST API

  • Addedyeet-seo CLI — check indexing status, submit pages, and run AEO audits from your terminal.
  • AddedMCP Server — expose SEO tools to AI agents in Cursor, Claude Desktop, and Windsurf.
  • AddedREST API v1 — programmatic access to indexing, inspection, AEO analysis, crawl budget, and trends. Bearer token auth with scoped API keys.
v0.4.0

Infrastructure Overhaul

  • ImprovedMigrated from Firebase/Firestore to PostgreSQL on Google Cloud SQL — faster queries, proper relational data, and SQL migrations.
  • ImprovedAutomated deployments via Cloud Build — push to main triggers build, migration, and PM2 restart on the VM.
  • ImprovedSecrets management via GCP Secret Manager instead of hardcoded .env files on the VM.
v0.3.0

AI Agent & Scheduler

  • AddedAutonomous AI indexing agent — Claude-powered agent that inspects URLs, decides which ones need indexing, and submits them automatically.
  • AddedCron scheduler — daily agent runs at 2am UTC with manual trigger support from the dashboard.
  • AddedAnti-spam safeguards — rate limiting and duplicate detection to prevent over-submission to Google's Indexing API.
v0.2.0

Rebrand to yeet.seo

  • ImprovedRebranded from index-booster to yeet.seo — new identity as an AI-native SEO agent.
  • FixedPM2 configuration for production deployments.
  • FixedRemoved legacy Firestore imports from API routes.
v0.1.0

Initial Release

  • AddedGoogle Search Console integration — connect properties and view indexing status.
  • AddedManual URL submission for indexing via the Google Indexing API.
  • AddedDashboard with per-URL status tracking, submission history, and crawl dates.
  • AddedGoogle OAuth sign-in with session management.

Frequently Asked Questions

What is yeet.seo?

yeet.seo is an API-first SEO toolkit for developers and AI agents. It auto-indexes pages via Google's Indexing API, analyzes AEO (Answer Engine Optimization) readiness, and provides tools via CLI, REST API, MCP server, and GitHub Actions.

What is Vibe Fix in yeet.seo?

Vibe Fix is a feature that generates AI-agent-ready prompts for each SEO insight. One click copies a detailed prompt with exact code changes, affected URLs, and file paths — paste it directly into Cursor or Claude Code to fix SEO issues using AI.

How often is yeet.seo updated?

yeet.seo ships updates frequently. This changelog tracks every feature, improvement, and fix. Major releases include new integrations, AI capabilities, and developer tools.