A hands-on field guide to Dalga’s Awesome YouTube Bot (AYB). The desktop app opens browsers or Android emulators, watches videos, and can like, comment, or subscribe with real Google accounts. This guide combines the product docs with transcripts from every AYB tutorial on Dalga’s channel.
1 · What It Is & How It Works
Awesome YouTube Bot is a YouTube optimization tool that mimics real human viewing. Instead of blasting fake views through an API, it opens a real browser (or an Android emulator), navigates to your video the way a person would (via YouTube search, suggested videos, Google, direct, or an embed), watches it, scrolls, changes quality/speed, and can like, comment, and subscribe with logged-in Google accounts.
vs. SMM panels · Why “real browser” matters
SMM panels push 100k views in hours via unofficial APIs, with no browser, no user, no interaction. YouTube’s AI spots and strips that. AYB is slower (it can’t open 100k browsers at once) but each view is a real watch with a unique fingerprint, so views look organic and stick.
The levers · What it can move
Views & watch-time, likes, comments, subscriptions, live-stream views, and crucially keyword CTR: search your keyword, scroll, find your video, click, watch to the end. That search-find-click pattern is the YouTube ranking signal.
The five view sources
| Source | What it does | Ranking value |
|---|---|---|
| YouTube Search | Searches your keyword on YouTube, scrolls, finds your video, clicks. This is YouTube CTR manipulation. | Highest. Feeds keyword ranking. |
| Suggested | Reaches your video through the suggested/recommended sidebar. | High. Signals relevance to other videos. |
| Google Search | Searches Google, then watches via the result (your URL must be indexed for that keyword). | Medium. External discovery. |
| Direct | Goes straight to the video URL. | Low. No discovery signal. |
| Embedded | Plays via an embed on a page. | Low/contextual. |
Source · "Awesome YouTube Bot" (FAQ walkthrough)
“AYB really mimics actual human behavior… it opens the browser, watches the video and interacts with it, scrolls through the page like a real user, changes the playback speed, quality and much more. Each browser the bot opens has a completely unique footprint… YouTube search will search your keyword on YouTube and watch it through that; suggested videos will generate traffic as the suggested videos; Google will search your keyword on Google.”
2 · Install & Requirements
Windows desktop software, like its sibling. For mobile views it also drives the LD Player Android emulator (MEmu also supported).
| Requirement | Detail |
|---|---|
| OS | Windows 10 / 11, or Server 2016 R2+. |
| RAM / CPU | 8 GB minimum, 4-core ~2600 MHz. More if you run many threads/emulators. |
| Framework | .NET Framework 4.7.2 or higher. |
| Browser | Latest Google Chrome for desktop views. |
| For mobile views | LD Player (the AYB installer can install it). Enable Intel virtualization (VT) in BIOS, usually on by default. |
| Version | V1.5.0 (Mar 2026) added 2captcha ReCaptcha solving + AI improvements. License = 2 devices. |
Source · "Android Phone Emulation in AYB" + product page
“First up we’ll set up LD Player… make sure Intel’s virtualization technology is enabled in your BIOS… AYB and LD Player support multi-threading… choose the emulator here, not Firefox, in order for the bot to run LD Player instead of the normal browser.”
3 · The Interface in 2 Minutes
AYB is built around a job list. You add jobs (each is a video + an action), set properties on the left, and hit Start.
| Area | What it’s for |
|---|---|
| Job panel | The tasks you’ve added (view / like / comment / subscribe / random act). Status + progress bar below, Start at top. |
| Import buttons | Import campaign (add a job), import accounts, import comments, import jobs from a text file. Plus a Loop function. |
| Left settings | IP method, threads/multi-threading, external browser (AdsPower), emulator vs Firefox, AI configuration. |
| Per-campaign accounts/comments | As of v1.4.2 each campaign can use its own account set and comment list. Press Edit job for changes to take effect. Cookies are stored after first login and restored on re-import. |
| Start individual | Start one campaign on its own via the campaign menu (bottom-right). Handy with many campaigns. |
Source · "Awesome YouTube Bot" + "New Recent changes of AYB"
“Now you can have different accounts and different comments for each campaign… just don’t forget to press the edit job button for your changes to take effect… you can start your desired campaign separately… press Start from the campaign menu in the bottom right corner.”
4 · Accounts & IPs: Get These Two Right
Almost every “it didn’t work / my views dropped” problem traces to one of these two. Nail them before anything else.
Google accounts
- Views: not strictly required, but accounts prevent drops. Per the latest YouTube changes, accounts are effectively needed for views to stick. Around 30 to 50 good accounts can support thousands of views.
- Likes / comments / subscribes: mandatory. You can’t engage as a guest. You need roughly one account per like/sub (1,000 likes is about 1,000 accounts).
- Quality > quantity. Aged accounts with an avatar and watch history massively outperform fresh ones. Low-quality accounts are the #1 cause of drops.
- Warm up weak accounts with a standalone Random Act job (Section 6). Let the bot watch random videos on different IPs to build history before you use them for real.
- Cookies are stored after the first successful login, so the bot doesn’t re-login every run. 2FA accounts are supported via a secret key; branded accounts (one Google account, many channels) via a
branded:prefix.
email@gmail.com:password
email2@gmail.com:password2|recovery@email.com
# 2FA account — add the 32-digit secret key after the password
email3@gmail.com:password3|recovery@email.com|JBSWY3DPEHPK3PXP
# branded account (one login, multiple channels)
branded:email4@gmail.com:password4IP changing
Same toolkit philosophy as ATB: residential IPs are what keep views alive. Blacklisted/free/datacenter proxies are the other big drop cause.
| Method | Notes |
|---|---|
| Proxies | HTTP(S)/SOCKS, multiple IPs at once = more views in less time. Supports rotating proxies via API link. Residential strongly preferred. |
| VPN | NordVPN (near-automatic with OpenVPN creds) or manual OpenVPN configs. Only one IP at a time, cheaper but caps throughput. |
| Android phone | Phone SIM data for residential mobile IPs (one IP at a time). |
| Bridge / PPPoE | ISP reconnect for free residential IPs where supported. |
IP:Port
User:Pass@IP:Port
# rotating — separate the refresh URL with a pipe
User:Pass@IP:Port|https://your-rotate-api-linkSource · "Awesome YouTube Bot", "Recent changes", "Branded Account Support"
“Regarding views we need [accounts] to prevent our views from getting dropped… for likes and subscribers we obviously need them… for a thousand likes we will need a thousand Google accounts… a two-year-old account with an avatar and some watching history is way better than a freshly created account.” · 2FA: “add the 32-digit two-factor authentication secret key after your password.” · Branded: “add the keyword branded followed by a colon before your account entry.”
5 · Your First View Campaign, Step by Step
1. Import accounts & set your IP
Import your Google accounts (Section 4) and configure proxies/VPN on the left. For views that stick, tick “use accounts” and use residential IPs.
2. Add a View job
Import campaign → action View. Enter a keyword (your video’s title works) and the video URL (trim to just the YouTube domain + video ID, drop extra parameters).
3. Pick the view source
Choose YouTube Search for keyword CTR (bot searches → scrolls → finds → clicks), or Suggested / Google / Direct / Embedded. YouTube Search and Suggested carry the most ranking weight.
4. Set watch time, quality & speed
Set a watch-time range (e.g. 60-90s). Watch at least 30-40 seconds, since YouTube ignores shorter views. Quality/speed can be random for realism; “default” picks data-saver to save proxy bandwidth.
5. Set count & (optionally) Random Act
Set how many views. Enable Random Act so the bot does natural activity (recommended/trending/short videos, random keyword searches, like/dislike) before the real view. Far more human.
6. Choose browser or emulator, then Start
Desktop = Chrome/Firefox; mobile = pick emulator (LD Player). Enable multi-threading to run several at once. Hit Start and watch the job panel / progress.
# keyword|URL : action : viewtime-or-commentlist : count
my video title|https://youtu.be/VIDEOID:view:60-90:15
https://youtu.be/VIDEOID:like::2
https://youtu.be/VIDEOID:comment:MyCommentList:5
# no keyword? skip the pipe and start with the URLSource · "Awesome YouTube Bot" (jobs + text-file import) + "Android Phone Emulation"
“A keyword should be written — it can be the name of your video… add the URL and set the duration… you have a few action options like view, like, random act, comment and subscribe, and the format is keyword | URL : action : view time or comment list name : count.” · “You must watch the video for at least 30 seconds… viewing a video less than 30 to 40 seconds, YouTube ignores this kind of views.”
6 · Likes, Comments, Subs & Nested Campaigns
Comments + spintax
Build a named comment list (one comment per line), then add a Comment job pointing at the video + list, and the bot picks one at random. Use spintax to multiply variety: {cool|interesting|amazing} makes the bot pick one word each time, so comments don’t repeat.
This is a really {cool|interesting|amazing} video, {thanks|appreciate it}!Random Act (warm-up & realism)
Random Act makes the bot do natural things: browse recommended/trending/short videos, run random keyword searches (spintax supported), like/dislike. It runs either around a real job or as a standalone job. The standalone version is the go-to for warming up Google accounts: run it for hours on different IPs to build watch history before using those accounts for real engagement.
Nested campaigns (the realistic combo)
Real fans don’t just view. They watch, then like, maybe subscribe, maybe comment, all in one sitting. Nested campaigns chain those into one browser session: View 20-30s → Like → View another minute → Subscribe → Comment. Set the count on the first job only (later jobs inherit it, so 100 first-job views give 100 likes). This single-session bundle is one of the strongest anti-detection patterns AYB offers.
Source · "Awesome YouTube Bot" (comments/spintax, random act, nesting) + integration video
“Nested campaigns… you can have a view job, then like the video, another view job, then a subscribe job, and finally a comment job — all of them in just one browser at the same time… define the number on the first job.” · “The standalone random act job is commonly used for warming up Google accounts.” · Integration: “I set a task to like the video 100% of the time… subscribe to the channel with a 35% probability.”
7 · Mobile Views & AI Mode
Mobile views via emulator
A large share of YouTube watch-time is mobile, so mobile views matter. AYB drives LD Player (or MEmu) to watch on a real Android environment, and each emulated device gets its own unique fingerprint. Set the view job up normally, then select emulator (not Firefox) before starting; enable VT in BIOS first. Multi-threading runs several emulator instances at once; pair with logged-in accounts to also prime them.
AI mode
Add an OpenAI (GPT-4o) API key in AI configuration and describe what the agent should do in plain language: open YouTube, handle cookie popups, watch realistically for N seconds, leave thoughtful comments, like, explore the channel, subscribe, hit the notification bell, do randomized actions. Run it before or after the main job with a timeout. It’s especially good for warming up new Gmail accounts with believable, varied activity.
Source · "Android Phone Emulation in AYB" + "AYB's AI is here"
“Choose the emulator here, not Firefox… each Android device created has its own unique fingerprint… link this with logged-in Google accounts; this setup is great for priming your Gmail accounts.” · AI: “instruct the AI agent to visit a YouTube video URL, handle cookie popups, watch realistically for set durations, leave authentic comments, like videos, subscribe and even activate notification bells… decide if the agent should run before or after the primary campaign action.”
8 · ★ The YouTube CTR & Ranking Playbook
YouTube ranks on satisfaction signals: did people search this, click it, watch a meaningful chunk, and engage? AYB lets you manufacture those, but the same realism rules decide whether it helps or backfires.
Watch-time and click-through beat raw view count. A real viewer searches, clicks, and stays. Engineer that, not a counter going up.
10,000 direct views that bounce at 5 seconds do nothing. 500 search-sourced views that watch 60%+ and occasionally like/sub can move a ranking.
1. Lead with keyword-CTR views
- Use the YouTube Search source on the keywords you want to rank for: search, scroll, find, click, watch. That’s the CTR signal that ties your video to a query.
- Target keywords you can realistically rank for (lower competition, where your video already appears a few rows down). CTR nudges; it doesn’t conjure rankings from nowhere.
- Mix in Suggested source to signal relevance to related videos.
2. Maximize watch-time honestly
- Always >30-40s, and ideally a real percentage of the video. Set watch-time as a range so it varies.
- Don’t watch 100% every time. Real audience retention is a curve, not a wall. Vary durations so some drop off early, some watch to the end.
3. Engage like humans, in proportion
- Use nested campaigns so the same session that watches also (sometimes) likes/subscribes/comments. That’s how real fans behave.
- Keep ratios believable: likes < views, subscribes « views, comments rare. A video with more likes than views screams manipulation.
- Spintax + varied comments so engagement never looks templated.
4. Quality of identity
- Aged, warmed accounts + residential IPs. This is the whole ballgame for views sticking. Warm new accounts with standalone Random Act before using them.
- Mix desktop + mobile (emulator) to mirror a real audience.
- Ramp gradually. A new upload that jumps to thousands of views in an hour is the classic fake-out. Drip it over days using the Loop function and scheduling.
5. Combine with Awesome Traffic Bot
Chain an ATB Google-search task with an AYB YouTube task in one flow so a single “person” finds your web result and your video and engages with both, the multi-surface footprint real interest produces. See the Traffic Bot guide
9 · Drops, Measuring & Risk
Why views drop (and how to stop it)
- Bad IPs: blacklisted, low-quality, or free proxies. Use residential. #1 cause.
- Weak / no Google accounts. Use aged accounts; warm fresh ones first.
- Watch time under 30-40s. YouTube ignores those views entirely.
Measuring impact
Watch YouTube Studio: views, average view duration / percentage, traffic sources (you want “YouTube search” and “Suggested” rising), impressions CTR, and ranking for your target keywords (search them and see where your video sits). Treat it as an experiment: baseline, change one variable, watch a week or two, keep a control video.
Source · "Awesome YouTube Bot" (drops + tips) + product page
“The most common reason [for drops] is the proxy — if we use blacklisted IPs, low-quality proxies or free proxies, YouTube can detect them… if our Google accounts are low quality, or worse you send views without any Google accounts, you will see drops… viewing a video less than 30 to 40 seconds, YouTube ignores this kind of views.” · Product page: “We strongly advise against using Awesome YouTube Bot on monetized videos.”
10 · Settings Cheat Sheet
Non-negotiables
- Watch time: > 30-40s, set as a range, vary it.
- IPs: residential only. Blacklisted/free proxies = drops.
- Accounts: aged + warmed. About 30 to 50 for thousands of views; roughly 1 per like/sub.
- View source: YouTube Search (CTR) + Suggested for ranking. Direct = filler.
Copy-ready formats
- Accounts:
email:password|recovery|2FA-secret· branded:branded:email:password - Jobs:
keyword|URL:action:time-or-list:count - Proxies:
User:Pass@IP:Port· rotating add|API - Spintax:
{cool|interesting|amazing}
The realistic-growth recipe
- Nested session: view 60s+ → sometimes like (varied %) → rarely subscribe/comment.
- Source = YouTube Search on a winnable keyword; add Suggested.
- Aged accounts + residential IPs; warm new accounts with standalone Random Act first.
- Mix desktop + emulator; ramp over days with Loop + scheduling.
- Keep likes < views, subs « views; track in YouTube Studio.
- Never on a monetized channel you care about.
11 · FAQ
Do I really need Google accounts?
For likes, comments and subscribes, yes, mandatory (you can’t engage as a guest). For views they’re not strictly required but they stop your views from dropping, so in practice you want them. Roughly one account per like/sub; about 30 to 50 good accounts can carry thousands of views.
Why are my views dropping?
Three usual suspects: blacklisted/free/low-quality proxies, weak or missing Google accounts, and watch time under 30-40 seconds. Fix all three: residential IPs, aged/warmed accounts, longer watch time.
Can it send 100,000 views in a few hours like an SMM panel?
No, and that’s the point. It opens real browsers, so throughput is bounded, but the views are real watches with unique fingerprints that look organic and stick, unlike API-blasted panel views that YouTube strips.
Should I use it on a monetized channel?
Dalga explicitly advises against it, and so does common sense. Artificial views on monetized content is invalid traffic that can get a channel demonetized or terminated. Keep it off anything monetized you care about.
How do I do mobile views?
Install LD Player (the AYB installer can do it), enable Intel VT in BIOS, set up your view job, then choose emulator instead of Firefox before starting. Multi-threading runs several emulator instances at once.
What are nested campaigns and why use them?
They chain view → like → subscribe → comment into one browser session, like a real fan. Set the count on the first job; later jobs inherit it. It’s one of the most natural, hardest-to-detect patterns AYB offers.
2FA or multi-channel accounts?
Both supported. For 2FA, append the 32-digit secret key after the password and the bot generates codes. For one Google account with multiple channels, prefix the entry with branded: and the bot switches between channels automatically.
12 · Sources
There is no written AYB manual; this is built from the product page and a transcript of every AYB tutorial on Dalga’s channel. Quotes link to the exact video.
Official
- Awesome YouTube Bot product page
- dalga.dev store / company
- docs.dalga.dev (ATB docs; AYB not yet covered)
Tutorial videos
- Awesome YouTube Bot full walkthrough
- Android Phone Emulation (mobile views)
- AYB’s AI agent
- Recent changes (per-campaign accounts, 2FA)
- Branded Account Support
- Integrate AYB with ATB
- Full DalgaDev channel
Awesome YouTube Bot Field Guide · Traffic Bot guide · Page glossary · samuelhuang.org
Independent reference for a third-party tool. Not affiliated with Dalga. Use within platform terms and the law.