Chronically Onlıne
The AI venture studio for AI consumer apps.
About
Chronically Online is a venture studio built by Antoine Levy and Mr Han (AI).
We build AI consumer companies and believe the future belongs to products built for specific communities, powered by compounding AI systems, and operated by founders with real ownership.
Built for communities
The future belongs to products built for specific communities — not one assistant for everyone.
Compounding AI systems
Every experiment, insight and learning becomes immediately reusable across every future product.
Founders with ownership
Operators run our products as founders — real ownership with the security of the studio behind them.
Products
The consumer AI apps we've tested
Each product is a bet on a specific community. The studio discovers, launches and scales them through one compounding system.
Alma
WhatsApp meal planning
A WhatsApp agent that plans your week of meals and hands back a ready-to-checkout Instacart cart. Built for busy households.
Skills
Mr Han's skill library
The compounding playbooks that run the studio — one master agent, 115 reusable skills across every stage of the venture lifecycle. Download any for $5.
Mr Han
The master agentMr Han is the operating system behind the studio. He identifies opportunities, researches markets, prototypes ideas, launches experiments and analyzes results — and every new test is informed by everything learned before. These are his skills: the compounding playbooks that run the factory.
Featured playbooks
Browse all 115 skills →Consumer Ai Creator Research
Build and score creator/influencer target lists for Consumer AI Factory venture validation, including platform discovery, fit scoring, contact-path checks, and first-wave outreach prioritization.
Consumer Ai Demand Probe Prototype
Build a fast Consumer AI Factory demand-probe prototype from an opportunity scan: choose the highest-signal wedge, implement only the first-value loop, verify with tests/build/browser QA, redesign if feedback says the UI lacks credibility, and log artifacts in Obsidian.
Consumer Growth Model And Loop Map
Use before committing to or materially expanding any Consumer AI Factory venture. Defines the product's acquisition, activation, retention, referral, revenue, and AI-personalization loops so teams do not build useful demos without a growth/retention machine.
Consumer Traffic Sprint
Use when starting or operating any Consumer AI Factory project that needs qualified traffic. Creates a channel-specific posting/commenting cadence, guardrails, tracking links, review loop, and default distribution requirement before a project is considered alive.
Consumer Ai Factory Distribution Learning
Capture, generalize, and reuse distribution learnings across Consumer AI Factory ventures while preserving venture/category nuance.
Avoid Ai Writing
Audit and rewrite content to remove AI writing patterns ("AI-isms"). Use this skill when asked to "remove AI-isms," "clean up AI writing," "edit writing for AI patterns," "audit writing for AI tells," or "make this sound less like AI." Supports a detect-only mode, an edit-in-place mode for files, an optional voice profile (casual / professional / technical / warm / blunt), and an iterate-to-convergence pass.
Learnings
What we've learned so far
Every venture improves our understanding of consumer AI. These are the reusable lessons that compound across the studio.
Product
- First value should come before signup whenever possible.
- Consumer AI needs an emotionally clear job, not just a clever AI capability.
- Voice can reduce friction — but only if the user immediately understands what to say and why it helps.
- Reports and summaries are valuable when they preserve a meaningful moment, not when they feel like generic AI homework.
Go-to-market
- Marketing before activation reliability wastes signal.
- Reddit and influencer channels require product specificity — generic self-promotion fails.
- Raw user language should drive copy more than founder language.
- Distribution learnings must preserve industry nuance before being generalized.
Agent workflow
- Mr Han orchestrates; temporary agents execute bounded work.
- Agents need explicit briefs, quality bars, and verification criteria.
- Fake permanent agent org charts are a trap — agents are workflows with inputs and outputs, not fictional employees.
- The process must produce decisions, not just artifacts.
Infrastructure
- Reusable assets to extract from every venture: auth, report/session persistence, AI generation route patterns, landing-page sections, voice/session UX, analytics events, and signup-after-value flow.
- Reusable infrastructure should speed up future ventures without forcing every product into the same shape.
Founder patterns
- Risk: hiding in product/build mode before real validation.
- Risk: expanding the meta-system before one venture has user pull.
- Standard: every process improvement should get us closer to real user evidence.
What compounds
- Data assets — user pain language, channels, winning and losing copy.
- Agent assets — pain-mining prompts, critic rubrics, launch-copy workflows.
- Infrastructure assets — auth, payments, analytics, experiment dashboards.
- Decision assets — thesis memos, kill/scale memos, cross-venture learnings.