✨ For You
Passive intelligence — I connect what I know about you (cards, trips, renewals, goals) into timely, useful nudges. Dismiss or snooze anything; I learn from it.
No insights yet. Add your cards & an upcoming trip below and hit "Refresh insights" — e.g. I'll flag lounge access for your flight.
👤 Your profile & wallet — the more I know, the more I can connect
{"benefits":["lounge","forex"],"activation":"DreamFolks via app"} · trip: set depart date + {"origin":"Bangalore","destination":"Surat"}
Nothing saved yet.
📊 Command center
Live runs, cost, providers, quality audits and your durable schedules — at a glance.
▶️ Running now
Nothing running.
🕑 Recent runs
💰 Usage by provider
✅ Quality — was your goal achieved?
Run "Verify quality" to audit recent runs.
🗓 System schedules — durable (stored in the DB, not app memory); toggle or run any built-in job
⏰ Scheduled research — deep/normal searches that run daily & on demand; every run is saved with history
SARVAM_API_KEY); the browser path above works keyless.🤖 Agents
Brain status & backend, the Claude-Code delegate, vision / media tools, and your roster of saved agents & skills.
Connect to load skills.
The team
Connect to load the roster.
🔎 Research
Deep, cited product / car / local research — comparison tables, claim-checks, price intelligence and booking links. Runs live or serves a fresh cached result.
🧩 Apps & widgets
Generate sandboxed mini-apps from a prompt, or pin live data widgets to a personal dashboard.
Your apps
No apps yet — build one above.
Preview
Select an app to preview.
📚 Knowledge
Your knowledge base, memory vault, learned playbooks and preferences — what the assistant knows about you and your world.
Namespaces
🧠 Agent memory vault — what the agent remembers (Obsidian-compatible PARA markdown, local & free)
🛠️ Skillbook — playbooks the agent learned from your tasks. Approve good ones (used more), edit, or remove.
No playbooks yet — they appear after the agent completes substantial tasks.
❤️ Preferences — what I've learned you like; I use these to rank suggestions.
🎯 Mission Control
Give me a complex goal — I'll think, plan a work-breakdown, spin up a team of specialists (researcher · trip-planner · deal-hunter · data-analyst · strategist), track each as a task, then rank + synthesize a decision-ready answer with booking links.
Recent missions
No missions yet.
Tasks
Drag a card between columns to change its status, or hit ▶ Run to have the agent complete it. Tick auto and the agent picks it up on its own.
Reminders
Deal watches
Profile & memories
Memories
🛡️ Approvals
Spend, booking, Claude-Code and app-write requests that need your explicit decision before they run.
📜 Activity
Every agent run with status, provider, tokens & latency — click a run to inspect its tool steps.
📈 Analytics
Product-event funnels (research → watch → alert, insight → acted), KPIs and the provider / cost mix.
⚙️ Settings
Backend, budgets, providers, channels, memory, connectors and security — everything configurable, in one place.
🤖 AI backend & models
Claude Code worker
/generated and cannot touch the host repo.💸 Budgets, memory & engine
🧠 Memory & engine
🔌 MCP config (advanced)
Raw settings
—