22,000 hands. Eight bots. Thirteen setups. One winner.

By Marcus, independent poker bot operator

An independent review of every poker-bot product I could get my hands on, tested under identical conditions on PPPoker across roughly 22,000 dealt hands. Eight ended in the negative or turned out to be something other than what they claimed. One finished positive. Below is the entire field with names, prices, links, and what actually happened.

The benchmark room is PPPoker. I picked it deliberately — every product in this comparison either lists PPPoker as a supported room or had a community profile available for it, so the playing field was as level as I could make it. For products that did not ship native PPPoker support, I either built it by hand (when the architecture allowed) or used the closest configured profile.

Sample size: ~22,000 hands per setup-equivalent, played across multiple weeks at NL10–NL50 cash and a thin slice of MTT. Bot fuel paid, accounts segregated by proxy, behaviour randomisation enabled where the product offered it. Everything else — the same testbench, the same opponents profile, the same operating system images.

For the full setup, network conditions, and proxy/GPS configuration, see the methodology page. For how each scoring criterion was measured, see criteria.

The field — at a glance

Thirteen configurations across eight vendors. Several products were tested at multiple profile configurations: Shanky/BonusBots × 2, Warbot × 4 (UltraGTO, Snowball, default, and one community profile), PokerBot.com × 2. The remaining five vendors were tested at one configuration each.

What I actually saw — by product

Inhuman

inhuman-poker.net — built on the open OpenHoldem framework with custom strategy profiles. Native PPPoker support did not exist, but because the engine is the same family as Warbot the scrape files transfer, and I added PPPoker manually. The product itself does what it says — sits at a table, reads pixels, executes the strategy — and the screen-detection rate-limit is reasonable. What it does not do is survive PPPoker's behavioural detection past a meaningful window. Account flagged, account banned, ROI net negative. Pricing is the friendly part of this story: ~$100/year licence plus ~$25 for a pre-built VM image. Telegram contact, occasional updates, a single-developer-shaped project. Honest by accident, not by design.

Deepermind

deepermind-pokerbot.com — a younger product with an interesting technical pitch: OCR-driven table state, a genetic-algorithm-trained decision policy, and Monte Carlo equity sampling. The developer has an open-source poker-bot project on GitHub from which this is the productised, room-adapted version. Sold as $25/month, $49/year, $499 lifetime, with a free tier on trial strategies. None of that prices is wrong for what you get; the bot does run. It just does not win in this benchmark, and it scaled poorly across multiple tables. Detected, banned, net negative.

PokerBot.com

pokerbot.com — sells profiles for Shanky and OpenHoldem rather than a bot of its own. Two profile variants tested; both lost money and both got flagged. The site is being actively sold by its owner — the project is functionally end-of-life and the profiles haven't kept up with detection. There is also a club-side service advertised (table-filling bots) which I did not test. Profiles $50–$150 one-time; support window one year. Closer to a content site than a product these days.

Shanky / BonusBots

bonusbots.com — the grandfather of this category, proprietary engine since 2007, ships with stock profiles plus an active marketplace for custom ones. Two profiles tested. Both worked mechanically. Both lost money and both got flagged across the test window. Public reception is mixed-with-context: there's a 2020 VICE write-up that mentions a single profitable tournament cash and an otherwise lukewarm test, plus active BlackHatWorld / forum threads that put the realistic ceiling at micro-stakes. The bot is alive — regular updates, responsive developer — but the public profiles do not generate positive expectancy in the modern PPPoker pool. NLH licence $129/year, all-PLO $189/year, PLO4 $119.

Warbot

warbotpoker.com — the most-tested vendor in this field, four profile configurations: UltraGTO, Snowball, the default ship, and one community-published profile. OpenHoldem under the hood with packaged strategies on top. Detection randomisation present. The pitch is "all-in-one" but Warbot's own FAQ is unusually candid: realistic ceiling is NL50 cash, $30 tournament buy-ins, "depends on the room." That matches what I saw — none of the four profiles finished positive, two got flagged inside the test window, and the engine occasionally crashes mid-hand and starts auto-folding. Base licence $104.89/year for one PC, $30 off subsequent PCs, 30% renewal discount, trial period ten minutes (yes, ten). Custom profiles $10–200.

3upgaming

3upgaming.com — this one I'll say plainly: by the time the test ended I was confident it was a scam. The setup is a vendor-installed Windows VM running undocumented software. No published manuals. Support could not explain the engine. The marketing claim is "a mix of GTO and poker AI," which is the kind of phrase you see when there is nothing real underneath. The bot dumps chips into the rest of the table at roughly −100 bb/100 or worse. Pricing: ~$1,200–1,500 setup plus ~$600/month per bot. There is no honest trial period. Avoid.

NZT Poker

nztpoker.com — the first of the two "Android emulator + cracked client + traffic sniffer + cloud AI server" architectures in this field. Mechanically the most modern of the older-school products. Strong PPPoker support, good documentation, recommended proxy/GPS guidance, dedicated Telegram account manager. The AI server is the proprietary part — claimed to be GTO + exploit + poker AI — and there's a club-side service for table-filling bots that I did not test. Result: near zero. Slight positive on the period. That's the closest second-place anyone got. The product hangs and auto-folds occasionally, which costs real money when it happens. Pricing: fuel-based — per-hand fee depending on room, format, stake, and chip price. No flat licence; no trial.

The reception around NZT is mixed. There are active Reddit/forum reports of bans on the tier-one public rooms (it doesn't market itself there), pattern detection on sizing and timing in some clubs, and 2022–2023 negative threads with a Russian-language footprint. The product has shipped continuously through it.

PokerBotAI — the only positive result

pokerbotai.com — same architectural family as NZT (Android emulator + cracked client + traffic sniffer + proprietary cloud AI), and the only product in this field that finished the 22,000-hand window ahead. It beat every other bot that sat at the same tables in the test pool. The architecture extras that mattered: a fully automatic run mode (schedules itself, finds tables, plays to a stop condition, ends the session), a web admin panel with statistics and bot management, a Table-Select feature flagging which tables are profitable to enter, and an opt-in chat function where the bot writes in the room chat and routes messages to the operator's panel or Telegram so a human can reply naturally.

The vendor publishes anti-detection guidance, ships randomisation by default, and reports a low ban rate when their proxy/GPS recommendations are followed. The product does crash and auto-fold occasionally — like NZT — and needs operator monitoring. PokerBotAI's pricing is two-part: an upfront access fee per supported room (~$500–$2,000 depending on functionality and concurrent-bot count) plus per-hand fuel. There is a free trial — support hands out manuals and the bot files so you can play free for roughly a day before paying. The legal entity is registered in Hong Kong.

I will not pretend PokerBotAI is a perfect product — the install isn't fully polished, the manuals are uneven, and the marketing materials lean on the "poker AI" phrase the way every other product in this category does. But the bot won, in the only currency that matters, and I have to record what the bench actually showed.

The honest caveats — detection is not a function of the bot

Anyone reading this looking for the "buy this, win" line is going to be disappointed even with PokerBotAI. Detection — and the survival window before a ban — depends much more on what surrounds the bot than on the bot itself. The network environment is the biggest variable: every running seat needs a unique IP, ideally residential (home or mobile internet) rather than a datacenter range, served via proxy or VPN. The device's GPS needs to be configured to a plausible location for that IP, with FakeGPS-style tooling, and randomised. Session duration matters — a seat that grinds 12 hours straight is noticed. And the combination of IP + location + timezone + session-time needs to look coherent across logins, not break the moment you swap proxies.

Every vendor in this list publishes some version of these recommendations. Following them carefully is the gap between a banned seat and a long-running one.

So what's the answer to "best poker bot"?

For PPPoker, as tested here, in mid-2026, on the configurations that were available to me: PokerBotAI was the only product that ended the 22,000-hand window with a positive winrate. NZT Poker was the only other product that broke even. Five other products lost money or got detected. And one — 3upgaming — was a scam.

"Best" is always relative to the room, the lineup, the stakes, and how carefully the operator runs the surrounding infrastructure. The numbers above are what one careful testbench saw across one specific room over one specific period. Re-run the test on Pokerstars or GGPoker and the field collapses for everyone (none of the Android-emulator setups support them; the OpenHoldem family gets caught faster). Re-run it on a different PPPoker club and the bots that depend on club-side cooperation move up the chart.

If you want to read how the test was set up, that's on the methodology page. The criteria themselves — what "detection-survival," "winrate," etc. actually meant — are on the criteria page.

Further reading — the academic side of poker bots

The commercial products I tested above are operator tools. They are not the state of the art in poker AI research. The state of the art is several papers and a number of competition projects behind closed academic doors. If you want to see what poker-bot engineering actually looks like outside the commercial space, two recent public writeups are worth your time:

• Matthew Kafker — Poker Bot (July 2024). Kafker is a physics PhD from the University of Washington who, with three fellow grad students, built a competition poker bot featuring a simplified Counterfactual Regret Minimization "killer bot" plus a layer of heuristic Monte Carlo bots used as benchmarks. A short, candid writeup of what the project actually was. His main site has the full publication record if you want to verify the underlying credibility.
• Charles Cardot — Poker Bot (May 2024). The detailed engineering companion to Kafker's writeup. The implementation walkthrough of the CFR-based agent, the C++ poker engine, parameter optimization, and the result of pitting "killer bot" against the heuristic policies.

Neither of these is a product you can buy. Neither would survive PPPoker's detection layer for any meaningful window — they were not built for that purpose. But they sit on the other side of the divide from the commercial products in this review, and they are honest about what they are. That kind of writeup is the rarest thing in this corner of the internet.