Daily Brief — 2026-05-27

> Generated: 2026-05-28 07:02 > Model: gemini-2.5-flash > Tokens: 43,901 in / 13,307 out > Cost: $0.0146 > Preset: small_team

---

The Day in Brief

The most pressing concern today is the AIOS CommandOS Bot, which has not exited successfully in 15 days, indicating a critical lapse in your core automation system. This yellow warning requires immediate attention, as your entire AIOS mission relies on these background processes. Without it, key data collection and automation tasks may be failing silently, undermining your strategic intelligence.

Your AI business revenue remains a critical gap. Stripe MRR is currently $89.75, with a slight negative balance, while your doTERRA income continues to provide essential runway. This highlights the urgent need to convert your extensive asset inventory into consistent revenue streams. The gap between your doTERRA income and AI income is the core problem to solve this year.

On the strategic front, there is significant Activator drift. While the Agent Opus Challenge is time-sensitive and the AI Asset Weekend Builder is your next product launch, several new high-priority projects have emerged. The GEO + Reputation Agency, the Testimonial Vault rebuild, the Coach Blog Website Template, and the Digital Home Website Service are all active, competing for your limited capacity. This rapid expansion risks diluting focus and delaying the monetization of your established priority stack.

Despite these challenges, your community continues to grow, with YouTube subscribers increasing and Skool rank improving. The AmplifAI Video Studio is live and generating revenue, and the Bookmark Vault is ready for distribution. The productization of the Digital Home Website Service and the GEO Agency presents significant opportunities, but careful management is needed to ensure they align with your "Revenue before Reinvention" rule. Today, focus on diagnosing the AIOS bot failure and re-aligning your active projects with your core priority stack.

Key Signals

⚠️ AIOS CommandOS Bot Stale: Your core AIOS bot has not exited successfully in 15 days, indicating a potential failure in background automation and data collection. This requires immediate investigation. 📊 Stripe MRR Low: Current Stripe MRR stands at $89.75, with a $-0.10 available balance, significantly below the $1,000+/mo Skool MRR target. 📌 Activator Drift High: Numerous new high-priority projects (GEO Agency, Digital Home Website Service, Testimonial Vault rebuild) are actively competing with your established 5-item priority stack, risking diluted focus. 🔥 Digital Home Website Service Productized: A new productized service, the Digital Home Website Service, has a proposed 3-tier pricing model and a streamlined 4-hour delivery plan per client. 💡 GEO + Reputation Agency Active: Phase 1 of the GEO + Reputation Agency is active, with tools installed and a strategy for free client audits to generate testimonials. ⚠️ Sara Dependencies Unresolved: The legal agreement for the Sara + Sondra agency and the build-out of AmplifAI Video Studio nurture sequences remain open items, posing a single point of failure risk. 🔥 Skool Community Growth: Your Amplif-AI Your Business community has 544 members (+33 in 7 days), and Skool rank improved by 503 spots in 7 days to #1621 overall. 💡 Blotato Competitor/Partner: Blotato, a social automation stack, offers a 5-skill system directly competing with or complementing your AI Marketing Team offer, warranting evaluation before launch. ✅ Bookmark Vault Ready: The first of 10 promised DFY apps, Bookmark Vault, is built and ready for distribution to founding Stars members and the wider community. ⚠️ Zapier Limit Hit: Your Zapier account hit its 100 tasks/month limit, holding 12 contacts until June 24, requiring a manual workaround or a Claude Code replacement. 📊 YouTube Subscribers Growing: Your YouTube channel now has 1,060 subscribers, up from 1,000, but videos published in the last 30 days are low at 3. 📌 Subscription Audit Incomplete: The 2026 subscription audit remains unaudited, with an estimated $500-700+/mo in spending, impacting cash flow visibility.

Metrics Analysis

Stripe Revenue (Direct Sales & Subscriptions): Your Stripe MRR is currently $89.75, generated from 18 active subscriptions. This is a significant increase from last week's reported $0.00 MRR, indicating new activity. Revenue MTD is $254.0, with 6 new subscriptions this month and 1 cancellation, resulting in a 5.6% churn rate. The income-snapshot.md shows the breakdown of these subscriptions, primarily from AmplifAI Video Studio tiers ($3/mo, $6/mo, $24/yr, $57/yr) and some $5/mo and $6/mo tiers. While any growth is positive, this figure is still far from the $1,000+/mo Skool MRR target, highlighting the need to scale your AI-powered offers. The negative Stripe balance of $-0.10 available, despite $13.67 pending, suggests tight cash flow that needs close monitoring.

Skool Community & Revenue: The Skool community, Amplif-AI Your Business, continues its strong growth trajectory, now at 544 members, an increase of 33 members in the last 7 days. Its overall rank improved by 503 spots to #1621, and #210 in the Tech category. This indicates effective awareness and discovery efforts, likely driven by your consistent Facebook content and weekend challenges. The Stars Premium MRR remains at ~$207/mo from 10 founding members, which is separate from the Stripe MRR. Skool affiliate earnings show a healthy lifetime total of $2,207.19, with a gain of $64.20 this week, demonstrating the value of your network. The challenge now is to convert this growing free community into paying Stars Premium members and other standalone product sales to close the revenue gap.

YouTube Channel (Awareness): Your YouTube channel has reached 1,060 subscribers, a positive indicator of growing audience reach. Total views are 61,593 across 163 videos. However, only 3 videos were published in the last 30 days, and views in the same period are low at 101. This suggests a falling behind on posting cadence, as noted in context/current-data.md. YouTube is a critical platform for evergreen content and discovery, so a consistent publishing schedule is essential to leverage this growth and drive traffic to your community. The low views for the last 30 days, despite subscriber growth, indicates that recent content may not be resonating or that the posting frequency is too low to maintain momentum.

Overall Impact: The data reveals a business with strong community engagement and a solid financial runway from doTERRA, but the AI business revenue is still in its nascent stages. The growth in Skool members and YouTube subscribers shows your content and brand are attracting attention. However, the low Stripe MRR and the warning about the AIOS CommandOS Bot are critical operational and financial signals. The Activator drift, with many new projects, could further delay focused monetization efforts on the core AI products. The immediate priority must be to stabilize the AIOS system and strategically convert community engagement into recurring revenue, focusing on the established priority stack.

Strategic Recommendations

1. Concerns:

• AIOS CommandOS Bot Failure: The 15-day stale status of your AIOS CommandOS Bot is a critical system health issue. This bot is likely responsible for essential background tasks, data collection, and automation that underpin your entire intelligence operation. Its failure means you may be operating with incomplete or outdated information, and critical automations could be silently failing. This directly impacts your "Away-From-Desk Autonomy" KPI.
• Activator Drift and Capacity Overload: Your "Active Priority Stack" lists 5 core items, but your "Dashboard" shows a proliferation of "🔴 Active" and "⭐ NEW" projects (e.g., GEO Agency, Digital Home Website Service, Testimonial Vault rebuild, Coach Blog Website Template). This is a textbook case of Activator drift, where new ideas are competing with existing, uncompleted priorities. This dilutes your focus, delays revenue generation from core products, and directly violates your "Finish before Starting" and "Revenue before Reinvention" rules.
• Low AI Business MRR & Cash Flow: Despite community growth, your Stripe MRR is only $89.75, and your available Stripe balance is negative. While doTERRA provides a runway, the AI business is not yet self-sustaining. The incomplete 2026 subscription audit further obscures your true financial picture and potential cost savings. This slow revenue growth, coupled with new project starts, puts pressure on your financial stability.

1. Opportunities:

• Productize Website Services (Digital Home & Coach Blog): The productization of the Digital Home Website Service and the Coach Blog Website Template offers a clear path to generate significant one-time and recurring revenue. With streamlined delivery and proposed pricing tiers, these can be quickly brought to market, leveraging your existing Base44 expertise. The Testimonial Vault rebuild as a flagship demo for the AI Asset Weekend Builder also fits this.
• Launch GEO + Reputation Agency: The GEO + Reputation Agency, with its installed tools and clear strategy for free audits leading to paid retainers, presents a high-ticket service opportunity. This directly aligns with your "AI Visibility Activator" identity and can generate substantial revenue quickly, especially with the identified South Florida health/aesthetics niche.
• Leverage Blotato for AI Marketing Team Strategy: The insights from Blotato's 5-skill social automation stack and Sabrina Ramonov's organic growth strategy are invaluable. Evaluate Blotato not just as a competitor, but as a potential affiliate partner or a source of inspiration to enhance your own AI Marketing Team offer before its public launch. The "Post Grader" feature is particularly worth replicating.

1. Quick SWOT:

• Strength: Sondra is a prolific builder with a rapidly growing community and a clear vision for AI-powered visibility.
• Weakness: Significant Activator drift is diluting focus and delaying the monetization of core AI products.
• Opportunity: Productizing existing app and website templates offers immediate high-ticket revenue potential.
• Threat: Critical AIOS system health warnings and low AI business MRR pose operational and financial risks.

Action Items

1. [Sondra] Investigate AIOS CommandOS Bot: Immediately check the AIOS CommandOS Bot process. Diagnose why it hasn't exited successfully in 15 days and restart/fix it. This is critical for system integrity.
2. [Sondra] Re-align Active Projects: Review the "🔴 Active" projects on your dashboard against your "Active Priority Stack." Pause or defer any new projects that do not directly contribute to the top 2-3 priorities (Agent Opus, AI Asset Weekend Builder, AI Marketing Team) to combat Activator drift.
3. [Sondra] Prioritize AI Asset Weekend Builder Launch: Focus on building the 3 flagship demo assets, setting up Stripe products, and completing the Base44 sales page. This is your next standalone public product and a key revenue driver.
4. [Sondra] Evaluate Blotato: Review the Blotato social automation stack. Decide if it's a competitor to outflank, a potential affiliate, or a source of features to integrate into your AI Marketing Team before its launch.
5. [Sondra] Formalize Digital Home & Coach Blog Pricing: Formally confirm the proposed pricing tiers for the Digital Home Website Service and the SEO Blog Website Template service. This unlocks the ability to market and sell these new productized offers.
6. [Sondra] Address Zapier Limit: Implement a manual workaround for the 12 contacts on hold due to the Zapier limit. Explore Claude Code as a long-term replacement to avoid future disruptions.
7. [Sondra] Complete 2026 Subscription Audit: Schedule a dedicated session to complete the 2026 subscription audit. Identify and cancel any non-ROI subscriptions to improve cash flow.
8. [Sara] Build AmplifAI Video Studio Nurture Sequences: Prioritize building the email nurture sequences (Standard onboarding, Elite onboarding, monthly drip) for the AmplifAI Video Studio. This is critical for member retention and upsells.
9. [Sondra] Draft Sara + Sondra Legal Agreement: Begin drafting the 50/50 written legal agreement for the agency work with Sara. This is a prerequisite before the agency becomes fully operational.
10. [Sondra] Post YouTube Video: You are falling behind on your YouTube posting cadence. Publish a video this week to maintain momentum and leverage your growing subscriber base.

Your Open Questions

• Sondra has NO laptop backup outside git for the aios-starter-kit/ folder. No OneDrive sync, no Obsidian paid sync, no Backblaze, no external cloud backup. Everything not committed to git lives on a single laptop. Risk: laptop failure = total loss of gitignored content (Notion sync output, YouTube transcripts, Granola meetings, Screenpipe logs, master .env, all secrets, the 33 leftover untracked files). Recommended fix: Backblaze ($9/mo, $108/yr) or OneDrive folder sync (free up to 5GB) this week.
• Answer: This is a critical vulnerability. You need to implement a robust backup solution immediately. Backblaze is a strong recommendation for comprehensive, automated laptop backup. Alternatively, configure OneDrive to sync your aios-starter-kit/ folder if it fits within the free tier limits. Do not defer this.
• Token expiration unknown — Sondra regenerated the "Claude Code" GitHub PAT on 2026-05-25 but the chosen expiration window (90 days / 1 year / no expiration) was not recorded. If anything other than "No expiration", needs a calendar reminder ~1 week before expiration to regenerate.
• Answer: Check your GitHub settings for the "Claude Code" Personal Access Token (PAT) to determine its expiration. If it has an expiration, set a calendar reminder for one week prior to regenerate it. If "No expiration" was selected, ensure it is stored securely.
• Misconfigured remote in PROJECTS/Pushten website/amplifai-template-studio/ — its origin remote points to AmplifAI-Your-Business/test-sondra-blog-website.git (same repo as the Sondra Blog Website folder). Either intentional clone or stale misconfiguration. Worth a quick check.
• Answer: This is a potential misconfiguration. You should perform a quick check to confirm if this was an intentional clone or a stale setting. If it's a misconfiguration, update the remote URL to the correct repository for amplifai-template-studio.
• Apps 2 and 3 not yet replicated — 69fbd5b9f9b68c868908f6ff and 6a00f7ca5edc423291aeb2d9 still need usePageSeo hook + all dashboard toggles + meta description fixes. Full replication checklist in PROJECTS/Sondra Blog Website/INTERNAL-MASTER-STATUS.md section 2.
• Answer: This is an open task for the Base44 Blog Template productization. Schedule a Claude Code session to complete the replication checklist for these two apps, ensuring consistent functionality and SEO.
• Prerender.io vs Vercel migration vs marketing reposition decision pending — three real paths to deliver full AI-search visibility. Research with pricing in master status doc section 5. Cost-to-buyer: Prerender.io adds $49/mo to every Base44 install.
• Answer: This is a key strategic decision for your SEO Blog Website Template. Review the research and pricing in PROJECTS/Sondra Blog Website/INTERNAL-MASTER-STATUS.md section 5. Weigh the cost implications for buyers against the benefits of full AI-search visibility. This requires a dedicated decision-making session.
• og:image default still Base44 logo in pre-React HTML — when anyone shares a URL, social preview shows Base44 logo, not branded image. Need to find where platform-wide default OG image is set. Untested.
• Answer: This impacts your brand's social media presence. You need to investigate where the platform-wide default OG image is set within Base44 or your app configuration and replace it with a branded image. This may require a Claude Code session to explore Base44 settings.
• 0 Content Types and 0 Niches defined on master — pSEO Generate batch system cannot run. Pre-built Listicle/Comparison/How-To/Best Of templates not yet inspected to confirm if their prompt_templates produce AEO/GEO content or SEO-only.
• Answer: This is a blocker for your pSEO content generation. You need to define Content Types and Niches on the master app. Also, inspect the pre-built templates to ensure they are configured for AEO/GEO content, not just traditional SEO. This requires a Claude Code session to configure the backend.
• Owner Playbook v2 acceptance — earlier rejected /implement run rebuilt the customer-facing Owner Playbook PDF with TOC + lookup table + Site Config crops. Current outputs/blog-template-owner-playbook.pdf is v2; v1 backup exists. Never explicitly accepted/rejected. Decision deferred.
• Answer: You need to formally review and accept or reject the Owner Playbook v2 PDF. Open it in Edge, verify clickable TOC and hyperlinks, and skim the Site Config crops. Provide feedback if adjustments are needed.
• Customer-facing Owner Playbook needs updating — to reflect verified state, buyer-required steps (Site Config completeness, custom domain, app rename, meta descriptions, all dashboard toggles, Prerender.io optional path), and honest Base44 limitations.
• Answer: Once v2 is accepted, schedule a Claude Code session to update the Owner Playbook to reflect the latest verified state, buyer-required steps, and transparently communicate Base44 limitations.
• Publish UX wall — STATUS dropdown → Published is non-obvious for non-technical buyers. Either document loudly OR send Base44 editor prompt to replace with [Save Draft] [Publish] button pair.
• Answer: This is a user experience issue for your buyers. Decide whether to clearly document the "STATUS dropdown → Published" process in the Owner Playbook or to use a Base44 editor prompt to create a more intuitive [Save Draft] [Publish] button pair.
• AI Write Post voice quality — produces generic AI slop with gendered framing and zero specifics. Buyers will refund or quietly stop using it after one test. Need voice-training intervention OR clear "starting draft only" positioning before selling.
• Answer: This is a critical quality issue that impacts buyer satisfaction. You must either implement a voice-training intervention for the AI Write Post feature to align with your brand voice or explicitly position its output as a "starting draft only" in your marketing and documentation.
• Visual review of v2 PDF — Sondra needs to open it in Edge, click 3-5 random Lookup Table rows to confirm hyperlinks jump correctly, and skim the Site Config crops to verify the auto-generated bands captured each field group. If a crop is off, adjust the percentages in scripts/crop_site_config.py and rebuild.
• Answer: This is a follow-up to the Owner Playbook v2. Please perform this visual review as soon as possible to ensure the document is functional and accurate.
• Length trim after navigation lands — defer until Sondra has actually used the v2 PDF. Navigation may make the page count feel less overwhelming on its own.
• Answer: This decision is deferred. Use the v2 PDF first to assess if the navigation improvements mitigate the perceived length.
• Document B (Client Guide) — still deferred per plan. Pick up only after Owner Playbook v2 is approved.
• Answer: This is deferred. Focus on the Owner Playbook v2 approval first.
• Site Config crops — auto-crop vs hand-crop — auto-crop bands shipped in v2 using straight horizontal % bands of the 3420x3855 siteconfig.png. If individual bands look cut off, hand-crop in Snipping Tool with the same filenames in PROJECTS/Sondra Blog Website/cheat-sheet/screenshots/site-config-crops/ and rebuild.
• Answer: During your visual review of the Owner Playbook v2, check the Site Config crops. If any appear cut off, use the Snipping Tool to hand-crop them and rebuild the PDF.
• Modal readability fix verification — .admin-white-card className added to AdminModal.jsx via Base44 prompt. Sondra reported "bad popup.png" of broken modal but the image is rejected by API (>2000px). Need her to verify visually or send a smaller crop after the most recent deploy finishes.
• Answer: Please visually verify the modal readability fix. If the issue persists, send a smaller crop of the "bad popup" image for further diagnosis.
• pSEO + indexing comprehensive fix deployment status — the big multi-feature prompt (auto-images during Generate batch, real Generate OG Images, real Rebuild Silo Links, /sitemap.xml route, IndexNow Submit All, Featured Image Picker, 2 new dashboard cards) was sent to all 3 apps but not yet verified end-to-end.
• Answer: This is a critical set of features for SEO. Schedule a Claude Code session to perform end-to-end verification of this deployment across all 3 apps.
• Master Base44 app selection — Sondra said "keep all 3 in sync" but eventually one becomes the production master for cloning to new buyers. Candidates: 6a08a00f73e8ade183de082d (test-sondra-blog-website, the one we've been customizing), 69fbd5b9f9b68c868908f6ff, 6a00f7ca5edc423291aeb2d9.
• Answer: You need to select one of these Base44 apps as the production master for cloning to new buyers. This decision should be made before actively selling the service.
• Pricing tier formal confirmation — proposed $1,500 + $97/mo (Launch), $3,500 + $197/mo (Standard), $7,500 + $497/mo (Premium). Sondra hasn't formally accepted yet.
• Answer: Formally confirm these pricing tiers for the SEO Blog Website Template service. This is a prerequisite for marketing and sales.
• First buyer / launch path — no committed prospect yet. Could be a doTERRA international partner, a Stars community member, a GEO Agency med-spa lead, or Sondra herself as the test build.
• Answer: Identify your first target buyer or decide if you will use yourself as the test build. This will inform your initial launch strategy.
• Pricing confirmation — Launch $1.5K + $97/mo · Standard $3.5K + $197/mo · Premium $7.5K + $497/mo. Accept or adjust?
• Answer: Formally confirm these pricing tiers for the Digital Home Website Service.
• Monthly management scope per tier — $97 = monitoring + minor edits · $197 = above + 1 article/mo + corpus tweaks · $497 = above + 4 articles/mo + performance review + email nurture. Accept or adjust?
• Answer: Formally confirm the monthly management scope for each tier of the Digital Home Website Service.
• Stripe model for e-commerce clients — manual payout for v1 OR set up Stripe Connect from day one? (Default: manual until a client actually needs e-commerce.)
• Answer: The default of manual payout until a client needs e-commerce is a pragmatic approach. Confirm this decision.
• Sara's role — Sondra solo through client #3 (default) or Sara co-delivers from #1?
• Answer: Confirm if you will handle the first three Digital Home Website Service clients solo or if Sara will co-deliver from client #1.
• First client target — doTERRA international partner (Business-in-a-Box prospects this week) / GEO Agency med-spa lead / Stars member? Pick one.
• Answer: Identify your first target client for the Digital Home Website Service.
• Wrangler CLI not authenticated locally — Cloudflare MCP handles most operations, but wrangler deploy and wrangler secret put need one-time wrangler login (browser flow). Run when convenient.
• Answer: Run wrangler login to authenticate the Wrangler CLI. This is needed for direct Cloudflare Worker deployments and secret management.
• Gemini output quality vs Claude — Sondra needs to read the first draft article in the dashboard ARTICLES tab and judge whether the writing matches her brand voice. If meh, tune the prompt with stronger voice anchoring or revert to Claude. Draft to review: "How to Use AI to Write a Weekly Newsletter in 15 Minutes."
• Answer: Please review the Gemini-generated draft article "How to Use AI to Write a Weekly Newsletter in 15 Minutes" in your dashboard. Provide feedback on its quality and brand voice to determine if prompt tuning or a revert to Claude is needed.
• Where was the original unauthorized Claude swap Sondra remembered? Digital Home wasn't it (always Claude from first commit). Candidates not yet checked: Pushten website, Amplif-AI Template Studio, Coach Blog Template, BraveBrand intelligence app.
• Answer: This requires a Claude Code session to investigate the other candidate projects for any unauthorized AI model swaps.
• Backend dashboard UI gaps — No "Add Topic" button on Pipeline page, no "Run Trend Scan" button, no "Write Article" button per card, sidebar items (LEADS / EMAIL / STATS / AGENTS / GRAPH) are placeholders for unbuilt modules, dark theme is too dark to read, sidebar scroll bug (overflow-hidden on body + no overflow-y-auto on aside in src/app/layout.tsx). Needs a dedicated UI improvement session.
• Answer: This requires a dedicated Claude Code session focused on UI improvements for the Digital Home backend dashboard.
• No POST handler on /api/content-calendar — Currently only GET exists. To enable the future Add Topic button, need to add a POST handler that accepts { title, search_query, target_keyword, intent_type, priority, status } and inserts into content_calendar.
• Answer: This is a backend development task. Schedule a Claude Code session to add the POST handler for /api/content-calendar to enable the "Add Topic" functionality.
• package-lock.json has 122 unrelated deletions in the backend repo working tree (not committed). Origin unclear — likely stale from a previous session. Verify or discard before next commit.
• Answer: Before your next commit to the backend repo, verify these deletions. If they are indeed unrelated and stale, discard them.
• Phase 4 or 5 next? Decision needed — Phase 4 (backend functions + Resend) vs Phase 5 (AI content via Claude API).
• Answer: Decide whether to prioritize Phase 4 (backend functions + Resend for emails) or Phase 5 (AI content generation via Claude API) for the Template-Blog-website migration.
• Phase 4 — Backend functions still needed: publishScheduledPosts (cron-driven), generateSitemap, weeklyReport. Will require Supabase Edge Functions + Resend for email.
• Answer: This is a development task for Phase 4.
• Phase 5 — AI content generation pending: AdminGenerate.jsx and AdminNewPost.jsx still call Base44's built-in AI. Needs Anthropic Claude API integration. VITE_ANTHROPIC_API_KEY env slot exists in .env.local, currently empty.
• Answer: This is a development task for Phase 5.
• Phase 6 — Media uploads + Supabase Storage not yet configured. Admin Media page will not work until this is built.
• Answer: This is a development task for Phase 6.
• GitHub auto-deploy to Vercel not connected — Vercel GitHub app needs org-level permission grant for AmplifAI-Your-Business. Currently deploying via vercel --prod CLI uploads. To fix: Vercel → template-blog-website → Settings → Git → Connect GitHub → grant org access.
• Answer: Grant Vercel GitHub app org-level permission for AmplifAI-Your-Business to enable auto-deploys for the Template-Blog-website.
• Custom domain for template-blog-website.vercel.app — not named, not configured.
• Answer: Decide on a custom domain for the Template-Blog-website and configure it in Vercel.
• Resend account for Phase 4 emails — reuse Digital Home's Resend or create separate? Decide before Phase 4.
• Answer: Decide whether to reuse the Digital Home's Resend account or create a separate one for the Template-Blog-website's Phase 4 emails.
• TweetGod Supabase restore timing — currently paused (pqawkiljbvsrgmbidxub). When does TweetGod bot go live and need its Supabase back?
• Answer: Decide on the launch timing for the TweetGod bot to determine when its Supabase project needs to be restored.
• Purchase links: what payment mechanism? Stripe direct links, GHL checkout, or ThriveCart? Decide before Phase 2 build — this determines how each "Buy Now" button is wired.
• Answer: Decide on the payment mechanism (Stripe direct links, GHL checkout, or ThriveCart) for the Amplif-AI Template Studio before proceeding with Phase 2 build.
• Custom domain for template studio? Currently live at amplifai-template-studio.vercel.app. Does it get templates.amplifaiyourbusiness.com or its own domain? 5-minute add in Vercel once decided.
• Answer: Decide on a custom domain for the Amplif-AI Template Studio.
• GitHub connection to Vercel not linked — deploy was done via local file upload (GitHub app permissions weren't granted for the private org repo). To enable auto-deploy on push: go to Vercel → amplifai-template-studio → Settings → Git → Connect GitHub → grant access to AmplifAI-Your-Business org.
• Answer: Grant Vercel GitHub app org-level permission for AmplifAI-Your-Business to enable auto-deploys for the Amplif-AI Template Studio.
• TweetGod Supabase paused — was paused to free a slot for blog-website-template project. Resume when TweetGod bot goes live: Supabase dashboard → TweetGod project → Restore.
• Answer: This is a duplicate of the previous TweetGod Supabase question. The answer remains the same: decide on the launch timing for the TweetGod bot.
• Trend scan uses Google News RSS — blocked by Cloudflare IPs — The /api/trend-scan endpoint fetches Google News RSS which blocks Cloudflare Worker IPs. Workaround: add article topics manually in the backend dashboard content calendar, then use Write Article. Long-term fix: swap RSS source for a non-blocked feed (e.g. Bing News RSS, or a paid news API).
• Answer: RESOLVED 2026-05-19: This was resolved by swapping to Bing News RSS. No further action needed here.
• Backend GitHub repo is private — digitalbeachgirl/digital-home-backend set to private 2026-05-20. Cloudflare CI is connected and auto-deploys on push.
• Answer: This is an informational update. No action needed.
• Frontend pages not yet branded — sondraverva.com shows Luke's neutral baseline. Use BraveBrand Prompt Builder to generate a redesign prompt, paste into Claude Code in the Digital Home frontend project.
• Answer: This is a branding task. Schedule a Claude Code session to brand the Digital Home frontend pages using the BraveBrand Prompt Builder.
• context/brand-detail.md not yet updated — This file is in .claudeignore (load on demand). Contains older Melissa ICA detail and brand psychology that predates the May 2026 AI Visibility Activator update. To fix: say "load brand-detail" then apply the same updates from the May 2026 ICA and Brand Identity documents.
• Answer: Schedule a Claude Code session to load context/brand-detail.md and update it with the latest AI Visibility Activator brand identity and Melissa ICA details.
• Service domain strategy — GEO/reputation service on sondraverva.com (blog + landing page) or a separate brand domain (e.g., aigeoscore.com)? Not decided.
• Answer: Decide on the domain strategy for your GEO/reputation service: integrate it into sondraverva.com or launch a separate brand domain.
• Self-serve app brand name — "GEO Score" / "AI Visibility Score" / Sondra-branded? Needed before Phase 4 build.
• Answer: Decide on the brand name for the self-serve GEO app before proceeding with Phase 4 build.
• Jonathan Mast's GEO app — from GPT Reseller Network coaching call 2026-05-18. What does it specifically do? Is it competitive with Zubair's tool or a reseller/distribution thing? Research before launch.
• Answer: Research Jonathan Mast's GEO app to understand its functionality and competitive positioning relative to Zubair's tool.
• Sara + Sondra legal agreement — agency work starts Sondra solo (Phase 1). Legal agreement needed before Sara co-delivers. Timing unknown — ask Sara.
• Answer: This is a duplicate of a previous action item. Prioritize drafting the legal agreement with Sara.
• Content corpus needs customization — DONE 2026-05-18. All 10 files populated from Obsidian brain + BraveBrand Intelligence App + Skool testimonial vault. Live in GitHub repo.
• Answer: RESOLVED 2026-05-18. No further action needed here.
• brand_context Supabase rows not yet added — 3 rows needed before AI content generation works: (1) category: cta / key: links — HTML CTA links; (2) category: identity / key: author — "Sondra Verva"; (3) category: content / key: image_style — visual style description. Add via Supabase Table Editor → brand_context → Insert row.
• Answer: Add these three rows to the brand_context table in Supabase via the Table Editor. This is necessary for AI content generation in Digital Home.
• Pages not yet branded — site shows Luke's neutral baseline. Homepage, About, Services still say "Digital Home Starter." Use BraveBrand Prompt Builder (prompt-builder-pink.vercel.app) to generate redesign prompt, then paste into Claude Code in the Digital Home project.
• Answer: This is a duplicate of a previous branding task.
• GitHub Actions need SITE_URL variable — daily-publish.yml and weekly-trends.yml won't run until backend is deployed and SITE_URL repo variable is set in GitHub Settings → Secrets → Variables tab.
• Answer: Set the SITE_URL repository variable in GitHub Settings → Secrets → Variables tab for the Digital Home backend repo.
• Backend setup pending — clone lukesbrave/digital-home-backend-starter into PROJECTS/DIGITAL HOME V1/Digital_home_backend_starter/. Shares the same digital-home-sondra Supabase project. Do in a SEPARATE Claude Code session with that folder as root (per Luke's CLAUDE.md).
• Answer: Schedule a separate Claude Code session to set up the Digital Home backend, cloning the starter repo and rooting the session in that folder.
• Resend domain verification pending — currently using onboarding@resend.dev as sender. Before going live, verify sondraverva.com in Resend dashboard and update RESEND_FROM_ADDRESS in .env.local.
• Answer: Verify sondraverva.com in your Resend dashboard and update RESEND_FROM_ADDRESS in .env.local before going live with Digital Home.
• RESEND_WEBHOOK_SECRET is placeholder — only needed for email webhook events. Not blocking.
• Answer: This is informational. No action needed unless you plan to implement email webhook events.
• Next.js workspace root warning — two package-lock.json files detected (AIOS root + project). Fix: add turbopack.root to next.config.ts. Not urgent.
• Answer: This is a minor technical fix. Address it in a future Claude Code session when convenient.
• Agent teams not yet enabled — requires manually adding "env": { "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1" } to ~/.claude/settings.json. Claude cannot edit that file (intentional deny rule). Do this before trying agent teams. Verify Anthropic billing first — teams cost 4x solo sessions.
• Answer: Manually add the environment variable to ~/.claude/settings.json to enable agent teams. Verify Anthropic billing costs before using this feature.
• Delete duplicate skill files from module-installs/ — module-installs/youTube_search_skill.md and module-installs/youtube_pipeline_skill.md are duplicates of already-installed skills. Safe to delete.
• Answer: Delete these duplicate skill files from module-installs/ to keep your workspace clean.
• Dashboard posting not yet confirmed end-to-end — Flask server had IndentationError (fixed), then failed to start. Next session: paste full terminal output to diagnose. Check that SKOOL_EMAIL and SKOOL_PASSWORD are in .env.
• Answer: Schedule a Claude Code session to diagnose the Flask server issue for dashboard posting. Ensure SKOOL_EMAIL and SKOOL_PASSWORD are correctly set in your .env file.
• Post as Sondra or as AI Admin? — The AI Admin account ("Sondra's AI Admin", sondraverva@gmail.com) is set up in the community. Current dashboard posts as Sondra's main account. Decide which is the right approach long-term.
• Answer: Decide whether the dashboard should post as your main Sondra account or the "Sondra's AI Admin" account for long-term consistency.
• Custom Frames vs full custom Obsidian plugin? — Sondra wants the dashboard INSIDE Obsidian with the folder tree on the left (like the video at https://youtu.be/VaGpWWiHXm8). Two paths: (1) Custom Frames community plugin embeds Streamlit inside Obsidian — 30 min, works today; (2) Full custom Obsidian plugin built with Claude Code — polished, 1-2 sessions. Decision needed at start of next session.
• Answer: Decide between using the Custom Frames community plugin for a quick embed or pursuing a full custom Obsidian plugin build for a more polished, integrated dashboard.
• INSTALL-AUTO-LAUNCH.bat not yet run — agentic-os-dashboard/INSTALL-AUTO-LAUNCH.bat needs right-click → Run as administrator. Until then, dashboard does not auto-launch at login.
• Answer: Run agentic-os-dashboard/INSTALL-AUTO-LAUNCH.bat as administrator to enable auto-launch of the dashboard at login.
• Notion Task Scheduler triggers not yet disabled — push_to_notion.py and notion_context.py are still registered in Task Scheduler. Need to disable (not delete) to complete Notion transition.
• Answer: Disable the push_to_notion.py and notion_context.py Task Scheduler triggers to complete your transition away from Notion as a primary data source.
• ffmpeg + Groq key ready (2026-05-17) — Both installed. ffmpeg PATH fixed. Groq key exists but may need to be placed in ~/.config/watch/.env. After Claude Code restart, run /watch on any YouTube URL — the skill's setup wizard will confirm or prompt for the key location.
• Answer: After restarting Claude Code, run /watch on a YouTube URL to confirm the Groq key location and full functionality of the claude-video skill.
• ⚠️ Anthropic Pricing Change (added 2026-05-16 — FLAG) — Programmatic (API) usage no longer counts toward $200/mo Claude Max plan — confirmed from Chase Hannegan's AIOS workshop May 14. This means Claude Code API calls (tool use, scripting, automations) may now be billed separately on top of the $200/mo subscription. Needs verification: check Anthropic billing dashboard and account settings to understand what does/doesn't count. May impact AIOS automation costs significantly. (meeting: 2026-05-14-claude-code-agentic-os-workshop)
• Answer: This is a critical financial flag. You must immediately check your Anthropic billing dashboard and account settings to understand the implications of this pricing change on your Claude Code API usage and overall AIOS automation costs.
• 10 Facebook photos unconfirmed — catalog has 10 real photos classified with defaults (curated: false). Need visual confirmation and proper expression/pose classification. (session 2026-05-15)
• Answer: Visually confirm and properly classify the expression/pose for the 10 Facebook photos in your thumbnail generator catalog.
• search_examples.py yt-dlp path bug — fails when called from inside the venv because yt-dlp isn't on the subprocess PATH. Fix: use shutil.which('yt-dlp') to resolve full path before calling. (session 2026-05-15)
• Answer: This is a technical bug fix. Schedule a Claude Code session to implement the shutil.which('yt-dlp') fix in search_examples.py.
• Gemini face likeness quality — clean-v1.png face is close but not exact. If likeness quality is insufficient, fallback option is to generate without face and composite Sondra's real photo in Canva. Evaluate after a few more generations. (session 2026-05-15)
• Answer: Evaluate the Gemini face likeness quality after a few more thumbnail generations. If it's insufficient, be prepared to use the fallback option of compositing your real photo in Canva.
• Scrape Creators API credits — $29/mo plan selected. Unknown how many credits per thumbnail research run. (session 2026-05-06)
• Answer: Monitor your Scrape Creators API usage to understand the credit consumption per thumbnail research run.
• Full YouTube workflow not yet built — Sondra wants a complete system: keyword/title research, description templates, competitor analysis, thumbnail generation, all connected. Scoped but not built. Start with competitor research next session.
• Answer: This is a major project. Begin by conducting competitor research for your YouTube project in the next session.
• Competitor research pending — find 3 AI educator channels at ~1k subscribers, pull top thumbnails, audit against Lyndon's 3-element framework. This informs the YouTube project design. Use Supadata for the pull.
• Answer: This is the first step for your YouTube project.
• Which Amplif-AI portal app to develop further? Two options built: (1) 69fdb3b330941425ae20e00b — navy blue theme, custom numbered pillars; (2) 6a020afa3d93fd72236a8dad — VCS-style clone with Amplif-AI content. Pick one direction. (session 2026-05-11)
• Answer: Decide which of the two Amplif-AI portal apps you will develop further as your primary community site.
• Lovable MCP not yet tested — Added https://mcp.lovable.dev to project config. Worth testing on a landing page build. (session 2026-05-11)
• Answer: Test the Lovable MCP on a landing page build to assess its functionality.
• Daily cross-post community URLs — Sondra has Skool communities she posts to daily but never provided the slugs. Need these to finish the cross-posting setup in scripts/skool_post.py. (session 2026-05-11)
• Answer: Provide the Skool community slugs for your daily cross-posting setup in scripts/skool_post.py.
• Weekend hourly posting queue — folder structure ready (posts/queue/), but no pre-written posts yet. Needs content before scheduler can be built. (session 2026-05-11)
• Answer: Create content for your weekend hourly posting queue (posts/queue/) before the scheduler can be built.
• Higgsfield CLI — installed but not authenticated. What will it be used for? Run higgsfield auth login to activate. (session 2026-05-11)
• Answer: Run higgsfield auth login to activate the Higgsfield CLI. Determine its intended use case for your workflow.
• Skool category IDs — only General feed (63ae75dad12a455abf1310b0c0084775) confirmed. Other category IDs not yet mapped. (session 2026-05-11)
• Answer: Map the remaining Skool category IDs to complete your Skool API integration.
• What is the actual current total Skool member count? Sondra confirmed +130 in 2 weeks and Tech rank #217 but did not state the absolute total. Was 505+ before — may now be 635+. Need to verify and update context/current-data.md. (session 2026-05-10)
• Answer: Verify the absolute total Skool member count and update context/current-data.md.
• Where did "Skool rank #1824" come from? Appeared in the daily brief and old compile log. Unverified — possibly overall Skool rank vs Tech category rank #217. Need to confirm which dashboard shows what. (session 2026-05-10)
• Answer: Confirm the source and meaning of "Skool rank #1824" to differentiate it from your Tech category rank.
• May 8–9 Screenpipe data permanently lost — PC was on (Littlebird confirms), Screenpipe loop died at 8:57am May 7, no recovery until May 10. No way to recover those 2 days. Note for reference. (session 2026-05-10)
• Answer: This is noted for reference. No action can be taken to recover the lost data.
• Littlebird.ai — worth trying? Sondra found this tool during research on Screenpipe privacy. Reads active window text only (no screenshots), cloud-processed (SOC 2), Mac + iPhone, free tier + $17/mo Plus. Could complement Screenpipe for meeting notes. Evaluate free tier for one week before deciding. (session 2026-05-07-2030)
• Answer: Evaluate the free tier of Littlebird.ai for one week to determine if it complements Screenpipe for meeting notes.
• OpenAI Privacy Filter model — upgrade later? The full 1.5B param model from HuggingFace would give 97% F1 accuracy on PII detection vs current regex. Deferred — revisit if regex misses anything important in practice. (session 2026-05-07-2030)
• Answer: This is deferred. Revisit if your current regex PII redaction proves insufficient in practice.
• Claude Memory Compiler (github: coleam00/claude-memory-compiler, 998 stars) — auto-captures sessions, extracts decisions/lessons via Claude Agent SDK, builds searchable knowledge base. Runs on Claude Max subscription. Could enhance or replace manual Obsidian brain compile. Worth evaluating.
• Answer: Evaluate Claude Memory Compiler to see if it can enhance or replace your manual Obsidian brain compile process.
• Context Engineering Intro (github: coleam00/context-engineering-intro, 13k stars) — different planning workflow: INITIAL.md → /generate-prp → /execute-prp. May overlap with /plan-feature we just installed. Compare before deciding.
• Answer: Compare Context Engineering Intro's planning workflow with your existing /plan-feature before deciding on its adoption.
• Open CoDesign first run — needs API key setup on first launch (select Local CLI mode). Not yet tested in practice.
• Answer: Perform the first run of Open CoDesign, setting up the API key in Local CLI mode, and test its practical application.
• Other affiliate programs (non-Skool) — may be paying direct to bank via ACH or PayPal. No automated tracking. Worth checking bank statements for unidentified income sources.
• Answer: Periodically check your bank statements for unidentified income sources from other non-Skool affiliate programs.
• Will Claude Code see retroactive edits to daily notes? — Sondra noted that if she fills in a daily note late (e.g., writing May 4 content while actually on May 5), Claude Code may not catch it unless it's also put in May 5. Needs a process clarification. (daily-notes/2026-05-04)
• Answer: Claude Code reads the entire Obsidian vault during /prime and when specifically directed. If you edit a past daily note, Claude will see the updated content on its next /prime or if you explicitly ask it to compile or analyze that specific file. For critical updates, ensure they are compiled into the wiki or explicitly brought to Claude's attention.
• Is Screenpipe data actually visible to Claude Code? — Sondra installed Screenpipe and can see data in Obsidian, but wasn't sure if Claude Code was reading it. (Answered partially — CLAUDE.md confirms screenpipe_to_obsidian.py writes summaries to Sondra Brain/raw/screenpipe/ — Claude can read these but doesn't auto-read them unless asked to compile) (daily-notes/2026-05-04, 2026-05-05)
• Answer: Yes, Claude Code can read Screenpipe data. The screenpipe_to_obsidian.py script writes summaries to Sondra Brain/raw/screenpipe/. Claude will process this data if you explicitly ask it to compile or analyze content from that directory, or if it's included in a broader /prime context load. It does not auto-read every new Screenpipe entry unless specifically instructed.
• Can YouTube transcripts be auto-pulled into Obsidian/Claude Code? — Sondra watches videos on her phone and TV (not captured by Screenpipe) and wants those transcripts captured automatically. No solution implemented yet. (daily-notes/2026-05-03)
• Answer: This is an open development task. A custom solution would be needed to capture transcripts from videos watched on external devices. This could involve a manual input system or a more complex integration if a watch history API is available for those devices.
• Is Gamma connected to Claude Code? — Sondra used Gamma for the May 6 workshop and wondered if she needs to connect it through Claude Code. Not yet confirmed. (daily-notes/2026-05-06)
• Answer: RESOLVED 2026-05-06: Gamma is already active via MCP (Model Control Panel) and its tools are available in Claude Code sessions. No further connection is needed.
• Can YouTube/Gmail/Microsoft accounts be connected to Claude Code? — Sondra asked about connecting these data sources. No implementation yet. (daily-notes/2026-05-06)
• Answer: RESOLVED 2026-05-15: Google Workspace CLI (gws) has been installed and authenticated, providing 13 skills for Gmail, Calendar, Drive, Docs, Sheets, Slides, and Chat. YouTube integration is partially covered by youtube_history.py and claude-video skills. Microsoft accounts are not yet integrated.
• What should I be teaching Stars members? — Workshop attendance was disappointing (only 4 paid for the May 6 workshop). Sondra is questioning what the right training topics are. No clear answer yet. (daily-notes/2026-05-06)
• Answer: This is a strategic question requiring community feedback. Consider surveying Stars members or analyzing engagement with past session replays to identify high-demand topics. The planned Claude Lab training is a good starting point, but tailor it to specific member needs.
• Move everything to GHL snapshot or keep paying $97/mo? — Sondra attended Travis's weekly session and noted she needs to make a final decision. Not yet resolved. (daily-notes/2026-05-06)
• Answer: This is a financial decision. You need to weigh the benefits and costs of moving everything to a GHL snapshot versus continuing to pay $97/mo. This requires a dedicated review of your GHL usage and needs.
• doTERRA international partner program — Someone from another country asked about partnering with the monthly drop. Sondra wants to explore this (set them up with GHL sub-accounts, Agent Opus affiliate). Not yet acted on. (daily-notes/2026-05-05)
• Answer: RESOLVED 2026-05-19: Agent Opus country availability has been confirmed, and prospects are identified. The "Biz in a Box" model for doTERRA international partners is moving forward. Schedule presentations ASAP.
• Base44 affiliate issue — Sondra mentioned she's not receiving affiliate money from Base44 referrals. Unresolved. (meetings/2026-03-24)
• Answer: This is a financial issue. You need to contact Base44 support to investigate why you are not receiving affiliate payments from referrals.
• Hermes Agents desktop — Julian Goldie mentioned this. Sondra wants to know if it's worth tracking. Not yet investigated. (daily-notes/2026-05-03)
• Answer: Investigate Hermes Agents desktop to determine its relevance and potential value for your business.
• GitHub MCP — Discussed with Claude as worth adding. Not yet confirmed as installed/active. (daily-notes/2026-05-03)
• Answer: RESOLVED 2026-05-17: GitHub MCP is permanently blocked. CLI is preferred.
• How to move OneNote content into Obsidian — Sondra still has content in OneNote she wants migrated. Not yet addressed. (daily-notes/2026-05-03)
• Answer: This is an open migration task. Schedule a session to plan and execute the migration of your OneNote content into Obsidian.
• NotebookLM notebooks and Claude Code — Should she open her NotebookLM notebooks and reference them through Claude Code? Best workflow unclear. (daily-notes/2026-05-03)
• Answer: RESOLVED 2026-05-18: NotebookLM is connected and queryable. The "AI GEO Guide" notebook is already being referenced. The NotebookLM Content Multiplier Workshop (May 20) further demonstrated integration.
• Agent Opus videos — Has downloaded some but not all. Needs to sort, match with descriptions/titles/keywords that were created in Manus. Not yet organized. (daily-notes/2026-05-03)
• Answer: Organize your downloaded Agent Opus videos, matching them with their descriptions, titles, and keywords from Manus. This is part of the Agent Opus Challenge.
• Thumbnail skill for Claude Code — Requires 30 photos of Sondra to train an avatar. Currently blocked on getting those photos together. (daily-notes/2026-05-06)
• Answer: RESOLVED 2026-05-15: The Thumbnail Generator is fully operational with 40 photos classified. No longer blocked.
• open-codesign (GitHub: OpenCoworkAI/open-codesign) — Flagged as a Claude Design alternative. 5K stars, MIT licensed, parametric sliders feature. Sondra saw this but hasn't evaluated it. (daily-notes/2026-05-06)
• Answer: RESOLVED 2026-05-07: Open CoDesign has been chosen as your primary design tool and installed.
• Courtney website pricing — Sondra fully rebuilt it in Base44 as cloneable agency template (May 8). Two people want to buy it. Price not yet decided. (daily-notes/2026-05-06, 2026-05-08)
• Answer: Decide on a price for the Courtney website cloneable agency template, as you have interested buyers.
• AI Clone Ad Challenge $47 Backstage Pass — Decided: No. Already paid $1k for AI Clone Ads program (Jordan + Jeff