I'm building Neotoma, the system of record for AI agents. The idea is a record, not recall. Without one, an agent works from whatever it happens to recall: part of the last session, a note its tool saved, or nothing. With one, it reads what was stored, from which source, and what is current now, so the agents you connect can defer to one answer. Neotoma is open source, runs locally, and connects to agents through MCP, a REST API, and a CLI.
The principle underneath is the same one that's driven all of my work: people should control their own data, memory, money, and digital infrastructure, not cede it to platforms that optimize for engagement over truth.
I work as a solo founder in Barcelona, operating with AI agents as a team rather than as tools. Every workflow, email, finance, content, and product, runs through a shared repo and a shared record. The agents follow the same playbook I do. That experience informs Ateles, an open-source project in development for organizing agents, tasks, and plans around Neotoma’s shared record.
Before this chapter, I spent nearly two decades building products across consumer web, crypto, and startups: writing and shipping at TechCrunch, co-founding Plancast (acquired by Active Network), co-founding KITE Solutions, advising and building with early-stage startups, leading user experience at Hiro for the Stacks blockchain, and running Leather at Trust Machines. You can see the full arc on my timeline.
Without a system of record, an agent works from whatever it happens to recall: part of the last session, a note its tool saved, or nothing. With one, it reads what was stored, from which source, and what is current now, so the agents you connect can defer to one answer.
My old agentic stack was one agent per session with me in the loop for every task. Ateles is the standing fleet that replaced it. A personal agent swarm defined in Neotoma itself, where agents, tasks, and plans are all entities, and every action is attributed, auditable, and reversible.
Three viral posts agree that durable state is what makes agent loops work, then hand the job to a markdown file. What loop state actually requires once you run more than one loop.
Agent-led evaluation tells a user whether Neotoma fits. The new issues subsystem lets the same agent file a bug or enhancement, exchange messages with the maintainer, and watch the change arrive, all through MCP, without a human ever opening a browser tab.
The first problem with running multiple agents is that they forget. The second is that they can't coordinate. A state layer that solves storage first and then adds signaling becomes a nervous system, not just memory.
When capable AI absorbs most artifact production in the execution middle, human leverage concentrates at foundation and review. Coherence stops living in hallway handoffs and starts living in explicit standards, cross-readable artifacts, and judgment you can defend at volume.
When many agents write to one memory, you need proof on every row: who signed, what they were allowed to do, and no shared bearer secrets. Neotoma integrates AAuth so each agent signs with its own key and writes through grant-scoped admission.
On the PM role, AI, and the strange decision to define your job around the skill that just got commoditized.
Fourteen years ago I wrote a postmortem for Plancast. Reading it again now, the failure modes weren't really about product mechanics; they were about missing personal AI agents, a proper substrate for them to write to, and a sovereign mesh between them. All three exist now, and the original mission becomes tractable in a way the feed era never could.
OpenClaw stores memory in markdown files. Neotoma v0.4.3 plugs in natively as a structured state layer underneath, giving OpenClaw agents provenance, entity resolution, and versioned history without replacing the agent.
Three independent AI agent platforms worth billions converged on plain text files for memory. The convergence validates the problem. The failure modes they share define what comes next.
Most teams bolt agent memory onto whatever database they already have. It works until two agents write to the same store. Then one bad write propagates at machine speed, triggering downstream actions before any human can intervene. The industry is heading toward a trust crisis that retrieval optimization won't fix.