After months of exploring photo-based connections and social features, we're returning to what Remember This was always meant to be: a personal context layer that lets AI actually work for you.
What We Learned from the Detour
For several months after December, we explored photo-based social features: turning photos into conversations, building relationship insights, creating memory cards to share with friends. It was compelling—photos are powerful, and we all want better ways to stay connected.
But as we built these features, we realized something: they only worked well with local AI models, and local models weren't powerful enough for what we actually needed.
The real vision for Remember This required powerful AI that could understand context, make connections, and act autonomously. But we couldn't rely on local-only AI, and we couldn't just send everything to the cloud without giving users control.
So we built the infrastructure first.
Building the Foundation: AI@YourService and ccremote
To make Remember This work the way it should, we needed two things:
1. AI@YourService - Let Users Fund Their Own AI Usage
Instead of us running cloud AI (expensive, privacy concerns, vendor lock-in), we built a system where you pay for your own AI usage directly. AI Wallet gives you transparent pay-as-you-go billing across multiple providers. You control the spending, you own the relationship with the AI providers.
This solves the business model problem: Remember This can stay free forever because we're not paying for your AI usage. You are—and you can see exactly what you're paying for.
2. ccremote.dev - Claude Code Remote Control and Orchestration
Claude Code is powerful, but it's meant for interactive development—not continuous background processing. We needed a way to programmatically control Claude Code sessions and make them work on queued tasks autonomously.
ccremote watches your queue (Markdown files in your vault), spawns Claude Code sessions,
feeds them tasks, and lets them organize your life in the background. It's the automation engine
that turns Claude Code from a development tool into a personal AI assistant.
With these two pieces in place, we could finally build what Remember This was always meant to be.
The Real Problem: AI Can't See Your Context
Here's why Siri, Alexa, and Google Assistant feel so limited:
- They don't know what you did yesterday
- They can't see the photo you just took
- They can't reference the voice memo you recorded this morning
- They don't know what you're working on or what matters to you
They're stateless question-answering systems. Every interaction starts from zero.
But what if your AI could see:
- The screenshot you took of that error message
- The voice memo where you said "remind me to follow up with Sarah"
- The photo of that receipt you need to file
- Your time log showing what you've been working on all week
That's the context layer that's been missing. And that's what Remember This builds.
The Original Vision: Autonomous AI, Not Another Chat Window
Remember This was always meant to solve this:
How do you give AI the context it needs to actually be useful in your personal life?
Not by chatting with it. By letting it watch your inputs and organize them automatically.
Here's how it works now:
- You take a photo, record a voice memo, or log some time (things you already do)
- The Mac app watches for new inputs and preprocesses them locally (transcription, metadata)
- It writes queue items as Markdown in your Obsidian vault
- Claude Code (via
ccremote) picks up the queue and processes it autonomously - Your inputs get transcribed, tagged, linked, and organized—without you asking
The result? A vault full of searchable, structured context that AI can actually use.
Why This Matters More Than Photo Sharing
Photo-based social features are fun. But they don't solve the core problem: AI can't act on what it can't see.
When your voice memo about "call the dentist" automatically becomes a tagged task linked to today's diary entry, that's not just convenient—it's fundamentally different from any other AI system you've used.
When Claude Code can see that you took three screenshots of error messages this morning, cross-reference them with your time log showing what project you're on, and automatically file them in the right project folder—that's when AI stops being a toy and becomes a tool.
This is the vision we're returning to. Not "AI for your photos" but "AI with actual context about your life."
What This Means for You
If you've been following along since December, here's what's changing:
- Less focus on social/sharing features. Those might come back later, but they're not the core.
- More focus on autonomous organization. The system should work in the background, not require your attention.
- Deeper Obsidian integration. Your vault is the single source of truth, not our UI.
- Better Claude Code automation. The
ccremoteengine is where the magic happens.
The tagline captures it perfectly: "Your Life in Markdown. Organized by Claude Code."
You create the inputs (photos, memos, logs). Markdown preserves them. Claude Code organizes and acts on them.
What's Next
Over the next few months, expect:
- Better queue processing and smarter automation
- More flexible CLAUDE.md instructions for customizing how your inputs get organized
- Improved voice memo transcription and photo context extraction
- Templates for common workflows (daily diary, project logs, receipt filing)
- Documentation on how to customize your setup for different use cases
This is the tool I personally need every day. It's how I remember things with ADHD. It's how I keep track of client work. It's how I make AI actually useful instead of just impressive.
And it's free forever, because I can't afford for it not to exist.
Try the Refocused Vision
If you want AI that actually knows what you're doing and can act on it autonomously—not just answer questions when you remember to ask—this is it.
Built for people who think in notes and photos, not dashboards.