Structure and meaning · Open source · human in the loop

AI in a scope.
You in the loop.

Eonity builds open-source tools that give content structure and meaning, so people and AI can work on it together. Each tool hands the AI a clear scope and a set of moves it can make inside it.

TYDAL for digital assets. TYKERF for web pages. TYSET for documents, coming soon.

Where the AI worksTOOL → SCOPE → MOVES
  • TYDALDigital assetsScopeA VaultRead, search and propose metadata
  • TYKERFWeb pagesScopeA block of a pageSplit, move, resize and rewrite
  • TYSETDocumentsScopeA section of a documentWrite, format and restructure
You review, approve and undoEvery move is validated and recorded
Three tools. One way of working. Concept illustration
Open sourceStructure & meaningScoped AI · MCPHuman in the loop

The Eonity tools

Different content.
The same way of working.

Each tool understands one kind of content and gives the AI only the moves that make sense for it. They share a design language, an open-source license and the TY in their names.

01Digital assets

The typed Digital Asset Layer.

Structure documents, images, datasets and code with schemas, enrich them with AI and share selected content through Vaults with people, applications and AI agents.

AI scope: a vault
02Web pages

Web pages, one block at a time.

Cut a page into named blocks and change it through validated, undoable moves. Kaity, its AI, builds a page from a sentence and then works on the one block you select.

AI scope: a block of a page
03DocumentsComing soon

Structured desktop publishing.

A document editor for structured publishing: text flows through the frames of master pages, and every format comes from a catalog. The AI writes and edits one section at a time through validated operations.

AI scope: a section of a document

How we build

Structure and meaning,
for people and AI.

AI is most useful when it knows what it is working on and where its work ends. Eonity tools make both explicit, so you can hand over real work and still keep the final say.

01

Structure first.

Schemas describe resources, a slicing tree describes a page, and a document has its own structure. Structure gives content meaning that people and AI can both read.

Schemas · slicing trees · documents
02

A scope for the AI.

Each tool gives the AI a defined area to work in and a set of named moves it may use there: a Vault, one block of a page, one part of a document.

Scoped tools through MCP
03

People in the loop.

The AI proposes and shows its work; you review, approve and undo. Every change is checked before it is applied, so nothing leaves its scope unnoticed.

Review · approval · undo

See it work

The AI proposes.
You decide.

Claude curating an exhibition through a TYDAL Vault, and Kaity building a TYKERF page block by block: the same scoped way of working.

TYDAL. Claude reads an exhibition through its Vault, proposes titles and descriptions, and asks before it writes them back.More about TYDAL
TYKERF. One sentence on the start screen, and Kaity builds a page about Lisbon, one named block at a time (2:48). Then you point at anything, leave her tasks, and give them all to her at once.See the tasks·More about TYKERF

Built in the open

Good knowledge
deserves open tools.

Every Eonity tool is open source under the Apache License 2.0. Run them yourself, inspect the code and contribute to their development.

They work with the AI you choose: connect Claude or another MCP client, and it gets the same scoped moves the tools use themselves.

Explore the repositories

About Eonity