Alex Lukianov

AI agents can't use phones yet. We're fixing that.
Real iOS devices, fully agentic. Any app, any scenario.

[Alex Lukianov @LigoAlexandr]

We built a way to automate iOS devices for any workflow.

One example: warming up TikTok accounts. It can automatically scroll content, engage with niche-specific posts (likes, comments, saves), etc.
You can launch dozens of accounts in any geo using either our cloud iPhones or your own devices.

You can manage everything from our dashboard or connect it to your own agent.

And this goes way beyond TikTok. We’ve built the infrastructure + vision models to automate any iOS app. Works with agents like Claude Code, Codex, OpenClaw, and others.

[Alex Lukianov @LigoAlexandr]

@AnxKhn @ycombinator Yes bro 🤝

[Alex Lukianov @LigoAlexandr]

Rejected twice by @ycombinator.

Top 10% both times.

Worth celebrating? 😂

[Alex Lukianov @LigoAlexandr]

@robertoclaim nope

[Alex Lukianov @LigoAlexandr]

received our rejection from a16z Speedrun

wondering if you guys got the same rejection message? I know YC often sends different rejection reasons depending on the application

[Alex Lukianov @LigoAlexandr]

@shrutikhobes yes, by email. We applied 1 day before the deadline. thanks, got it!

[Alex Lukianov @LigoAlexandr]

Hey Codex, open LinkedIn on my iPhone, find the first post in my feed related to AI development or management, analyze it with context, and leave a comment.

P.S. Everything is executed on a real iPhone connected to a Mac + Codex + BoostLayer API.

[Alex Lukianov @LigoAlexandr]

@TimHaldorsson Hey Tim! Would love to chat! We’re testing some solutions for AI GTM. I’m based in Picoas, but any place works for me

[Alex Lukianov @LigoAlexandr]

Hey Codex, open LinkedIn on my iPhone and scroll through the feed

[Alex Lukianov @LigoAlexandr]

@subramanya yes, that’s why security is one of our top priorities before scaling broadly

[Alex Lukianov @LigoAlexandr]

Some raw experiments with new apps 👀

BoostLayer + Claude Code + iPhone + ESP32 controller = your agent can execute any workflow on your iPhone.

[Alex Lukianov @LigoAlexandr]

#2/2 TikTok automation demo - fully autonomous TikTok posting on a real iOS device.

Your favorite agent + iPhone = automation for virtually any mobile workflow.

[Alex Lukianov @LigoAlexandr]

🧵 TikTok Farm Playbook — Part 15/20: Posting & distribution (why scheduling tools destroy reach and how to post natively at scale)

⬆️ Previously: Format lifecycle — 5 phases from early signal to death. Exit at 6-9 consecutive underperforms. Don't confuse format death with account death.

You've got warm accounts, tested content, and a format in expansion. Time to distribute.

And this is where people blow it by trying to be "efficient."

**Rule #1: NEVER use third-party scheduling tools**

Later, Hootsuite, Buffer — doesn't matter. TikTok reads upload metadata. Non-native uploads carry different signatures. The result: suppressed reach, sometimes immediately, sometimes slow decay.

The only safe method: post from the app.

**Rule #2: Vary your metadata**

TikTok cross-references accounts through:
- Reused captions
- Repeated hashtags
- Copy-pasted descriptions
- Identical posting times

If 5 accounts post at 2:00 PM with the same hashtags, that's a signal. Vary captions slightly. Randomize posting times by 10-30 minutes. Rotate hashtag sets.

**Rule #3: Never repost a deleted video**

TikTok fingerprints content and remembers deletions. If a video flopped and you deleted it, you can't just repost it.

Instead: mutate it. Different hook frame, different text overlay, slightly altered pacing. Same idea, different fingerprint. And repost on a different account first.

**Rule #4: Posting cadence**

- Baseline: 1-2 posts per day per account
- Aggressive max: 3 posts (only during expansion phase)
- Scale HORIZONTALLY (more accounts) not VERTICALLY (more posts per account)

If reach drops as volume rises → cut posting in half, rotate formats. TikTok punishes volume on single accounts hard.

⬇️ Next: Comment engineering — the hidden conversion engine that most operators completely ignore

🔁 Repost if useful. Follow for the full series.

[Alex Lukianov @LigoAlexandr]

@MikelCobian The problem is that your reach will likely get reduced because of that. We started the same way.

[Alex Lukianov @LigoAlexandr]

@DanKulkov been following this playbook for a while

[Alex Lukianov @LigoAlexandr]

@om_patel5 doesn't he have a AI UGC avatar yet? :)

[Alex Lukianov @LigoAlexandr]

@paulg Thanks for these words, Paul

[Alex Lukianov @LigoAlexandr]

@athcanft warming up TikTok and Inst accounts

[Alex Lukianov @LigoAlexandr]

based on what i usually see - a shadowban is when you’re getting <100 views for 5-7 days straight.

daily engagement matters a lot (not just posting). Interacting with content every day (scrolling, liking, commenting, etc.) seems to reduce the chances of getting shadowbanned.

btw we built an API to control iPhones and automate pretty much any workflow - including everyday account warm-up, posting content, and more.

[Alex Lukianov @LigoAlexandr]

@vitaliidodonov @github What if Stanley could perform any action in X directly on your iPhone?

We automate real iOS devices, which means you can automate virtually any workflow and unlock way more use cases for Stanley users.