The AI that reads your game chat - so a human doesn't have to read the worst of it.
A logo built from four letters gamers type after a match: GG WP - "good game, well played." The company wants online play to actually earn the phrase.
Here is a fact that sounds made up but isn't: in 1997, a 20-year-old named Dennis Fong won a Ferrari in a video game tournament. The car belonged to John Carmack, the co-founder of id Software, and the game was Quake. The Wall Street Journal called Fong "the Michael Jordan of video games." Guinness lists him as the world's first professional gamer. This is the person who now runs a company whose entire job is to make online games less unpleasant.
The company is GGWP, founded in 2020, and the idea is straightforward even if the execution is not. Multiplayer games generate an enormous river of chat and voice, and some meaningful fraction of that river is harassment, slurs, brigading, and the general creativity of people trying to be awful to strangers. Games have historically dealt with this in two bad ways: keyword filters that catch "idiot" but miss the genuinely coded stuff, and human moderators who burn out reading the worst messages humans send.
GGWP's pitch is that this is fundamentally a data problem before it is a policy problem. Its platform, Community Copilot, ingests text chat, voice, usernames, player reports and Discord activity, then uses machine learning to read them in context - the sarcasm, the regional slang, the workaround where someone spells a slur with a zero. It flags what matters, triages the reports, and hands humans only the calls that actually require judgment.
What is genuinely interesting about GGWP is not that it filters chat. Lots of companies filter chat. It is that GGWP treats reputation as a longitudinal thing. Instead of scoring a single bad match, it tracks a player's behavior over time - something closer to a credit score for how you treat other people online. That reframing quietly changes the incentives, because a one-off outburst is different from a pattern, and the system is built to tell them apart.
The other thing worth noticing is who agreed to fund this. GGWP's cap table reads like a who's-who of the people who built the modern internet's communities: Riot Games, Sony's innovation fund, the two founders of Twitch, the founder of YouTube, the CEO of Krafton. When the people responsible for the largest online communities on Earth all write checks to the same trust-and-safety startup, the reasonable inference is that they have looked at the problem from the inside and concluded it is not a feature you bolt on. It is infrastructure.
And infrastructure is the tell here. In 2023 GGWP did the counterintuitive thing and made its core platform free to use. For a venture-backed company this looks like leaving money on the table, until you remember that toxicity is a shared problem and the moat is not the paywall - it is the data. Give the tool away, learn from every community that adopts it, and get better at reading language than anyone who charges by the seat. Distribution beats extraction, at least when your product improves with volume.
"An AI platform built from gaming to keep communities safe across chat, voice, and UGC."— GGWP, on what it does
The end-to-end platform: proactive moderation across text, voice, reports and Discord, plus sentiment analysis to watch community health.
Context-aware, real-time text filtering across 20+ languages that catches adversarial workarounds keyword lists miss.
Monitors voice within the full context of the game and conversation, available across 12 languages.
Automatically sorts and prioritizes player reports so moderators spend time only where it counts.
Detects offensive usernames and the loopholes players use to slip past filters.
Long-term player reputation scoring and real-time sentiment analysis to guide product and community decisions.
"Thresh." World champion in Doom and Quake, Esports Hall of Famer, and by Guinness's count the first professional gamer. Has co-founded five gaming/SaaS companies with over $1B in exits, including Xfire and Lithium.
Founded Crunchyroll, the anime-streaming platform Sony acquired for over $1 billion. UC Berkeley EECS/Math, with a doctorate from Carnegie Mellon. Knows building communities at scale.
Former CTO at Cyence and Chief Data Scientist at Cray. Machine-learning and game-theory background with a PhD from UC Irvine - the data engine behind the platform.
| Round | Amount | Date | Lead / Notable investors |
|---|---|---|---|
| Seed | $12M | Mar 2022 | BITKRAFT Ventures, Makers Fund, Griffin Gaming, Sony Innovation Fund, Riot Games, Emmett Shear, Kevin Lin, Steve Chen, Pokimane |
| Series A | $15M | May 2026 | Headline Asia, Smilegate Investment, Korea Investment Partners |
Total disclosed funding ≈ $49M. Revenue and valuation figures are approximate / undisclosed.
Fong, Gao and Ng start the company to combat toxicity in online games with AI.
Launches with a BITKRAFT-led round and backers including Riot Games and Sony.
Opens a free tier of its AI moderation platform to democratize access.
Extends its moderation AI to media networks and fan platforms like Fandom.
Raises a round co-led by Headline Asia, Smilegate and Korea Investment Partners to localize across Japan and Asia.
It builds AI-powered moderation and community-analytics tools that detect and respond to toxic behavior across text chat, voice, usernames and Discord for games, media networks and fan communities.
Founded in 2020 by esports pioneer Dennis "Thresh" Fong (CEO), Crunchyroll co-founder Kun Gao (COO) and data scientist Dr. George Ng (CTO).
About $49M total, including a $12M seed round in 2022 and a $15M Series A in 2026.
Game studios, publishers and media platforms including Unity, Sony, Krafton, thatgamecompany and Fandom - 25+ partners across 100+ games.
GGWP uses context-aware AI and cultural intelligence to understand slang, sarcasm and coded harassment across 20+ languages, and tracks long-term player reputation rather than isolated incidents.