many people don’t know this, but before any of this I was writing Solana programs in Rust, benchmarking hot paths, shaving compute units, thinking about account layouts, transaction contention, and how much useful throughput you could actually squeeze out of the runtime. at one point I was genuinely trying to get hired onto the @solana_devs team. funny how the path bends. the show is only starting.

Anyone can now launch a token through $AMM infrastructure: amm-bot.org/launch every launch gets its own vault, autonomous liquidity agent, and liquidity book. 100% of the token’s creator fees become working capital for that token. the agent deploys that capital across LP positions where it can continuously earn. the LP fees generated by those positions are automatically allocated: creator fees → LP buckets → LP fees earned → 60% buys back $AMM + 40% buys the launched token every token launched through amm-bot.org becomes its own independent liquidity engine while contributing to $AMM, our platform token. pump.fun made launching a token incredibly easy, but the liquidity layer is still one of the weakest parts. too many coins end up with thin, fragmented liquidity and no real way for turning their own trading activity into deeper liquidity. I’m optimistic that this can change that. launches should be able to generate, deploy, and grow their own liquidity automatically. launch: amm-bot.org/launch explore: amm-bot.org/explore

how high is the ceiling when a token's liquidity is managed by an autonomous agent that keeps turning creator fees into new LP buckets, then uses the fees those buckets generate to buy the token back? more capital means more buckets. more buckets means more fee-producing liquidity. those fees keep feeding back into $AMM. powered by @MeteoraAG, @Raydium & @DeFiTuna.

the $AMM docs are now live. every part of the system is documented in one place: how creator fees enter the machine, how pools are discovered and filtered, how LP positions are opened, what single-sided liquidity means, how bins and ranges are selected, how principal is kept separate from LP income, how fees are harvested, and how those fees ultimately route back into $AMM buybacks. for developers, there is also a public read API that exposes the agent's live state directly. you can query open positions, pools in scope, deployed liquidity, harvested fees, completed cycles, and the rest of the machine's activity without scraping the site or relying on screenshots. if you're an LP nerd, or just trying to understand the process behind $AMM, start here: amm-bot.org/docs

the doors are opening very soon. 💧