Launch roadmap
One milestone matters most: opening the invite-only beta, with five founding teams on brains of their own. This page shows the road there, the gate the beta has to pass, and the plan to the public launch and WordCamp Europe. The order is firm; the dates are proposals the team may still move.
What every TomaTío brain promises the people who use it. It ships with the beta; export with the Portable Context Protocol (PCP) is what makes it true.
The rule that sets the date: PCP export ships before anyone outside the team puts real data in a private brain. That includes restore and import, because turning export on promises that a brain can take its own archives back.
From the Spanish «¡toma, tío!», “take that, dude”: the brain hands your agents the context they need. The mark is a tomatillo wrapped in green ribbons.
The critical path, now nearly done.
Nothing blocks this. A few trusted people get a brain of their own now, to shake out onboarding before the beta. Test data only until PCP export ships: nobody puts real private data in before they can take it out again.
Quietly, when the beta opens: GPL from day one, and honestly early. The loud launch waits for the “Open it up” gate.
Every item in the checklist below is done. The highlights:
The major milestone: real teams using brains of their own.
All of these must be true before outside teams get a private brain.
| What | Status |
|---|---|
| Stable ids for people, agents and scopes | Done |
| Export a whole brain: archive, manifest, admin page and command line | Done |
| Shared documents and handoffs in exports | Done |
| Erasure receipts and purge, as PCP defines them | Done |
| Restore into an empty brain | Done |
| Import from another brain, as proposals a person reviews | Done |
| Every build checks archives against the PCP conformance suite | Done |
| A round trip between two test brains | In progress |
| Export switched on for everyone; the Bill of Rights in the README | Planned |
| What | Status |
|---|---|
| The name: TomaTío | Done |
| Domains, the mark, favicons and a public home at tomatio.ai | Done |
| The code public on GitHub | In progress |
| What | Status |
|---|---|
| A new brain starts with a team scope, so what agents write has somewhere to go next release | Fixed |
| No silent data loss: a failed save is never reported as saved next release | Fixed |
| User documentation: what it is, getting started, connecting each AI (Claude first, then ChatGPT, Cursor and VS Code), sharing, your data and leaving, troubleshooting | Planned |
| A beta note: what is stored, where, who can see it, and how to leave | Planned |
| Uptime alerts and backups for every brain | Planned |
| A 2-minute walkthrough video | Planned |
| Share links that open without a sign-in | Done |
TomaTío reaches teams through the hosting companies they already use, and PCP matters only if other tools speak it. Both have to be under way before TomaTío is offered widely, so adoption is one of the checks in the “Open it up” gate.
We invite hosting companies to join as co-contributors, not customers: it is how the WordPress community works, and a host that builds on TomaTío adds to it rather than competing with it.
An archive from a second, independent PCP implementation imports into a TomaTío brain, and back. Two implementations exchanging archives is the strongest proof that portability is real.
Goal: one or two hosting companies already offering TomaTío, with PCP export working in both products, and our work shared with the community on the WordPress AI ecosystem.
Are brains still active after two weeks? Memories written and approved, roles switched, agents' proposals accepted. What we learn shapes the conversations with hosting companies.
Start a brain on a laptop with no setup (WordPress Studio or Playground), then move it to hosting with PCP export and import. Not planned; revisit after the beta.
Two weeks after the beta opens, the team checks six things. The public launch goes ahead only if all six hold:
Opt-in per scope: private, members only or public. Only approved knowledge is published, never raw memory, as cross-linked pages with an index, a changelog, citations and provenance. Memories are the sources, approved knowledge is the wiki, procedures are the rules.