An ecosystem of products assembled from components within one metamodel.
Specialised editors for websites, app screens, images and data share a model of objects and relationships.
AI agents in Init and Frame create and edit pages and diagrams, save the result and let you undo changes. A shared core built on Bifrost connects OpenAI models and provides an API for other applications.
A collection of websites for different tasks, from shopping and trip planning to events
Open each site to explore how its content and user journey shape the interface.
Tools for working with AI agents in remote development environments
The agent runs on a Mac while builds, tests and containers run on a server, with separate Coder workspaces for each project.
A visual workspace for preparing an agent’s brief
Arrange documents and their relationships, select the context a task needs, and hand them over together.
A workspace for tasks carried out by AI agents
It brings together work streams, selected context and external session activity, keeping project changes as proposals to review.
A collaborative form builder with offline editing.
Each publication fixes a version of the form, so later edits do not change the questions people are already answering. Responses can be exported as CSV.
Private event planning through Telegram and an installable web app.
Invitations, attendance replies and schedule confirmation keep the decisions about a gathering together.
A trip planner for comparing route options.
A map and timeline connect stops and journeys, with cost and travel-time comparisons to help choose a plan. Routes can be assembled manually or imported.
A desktop character that reminds you to blink, rest and stretch while working at a computer.
Delegate a task to an independent AI agent from your main working session
It retains progress and results, lets you continue the conversation, and returns proposed changes for review.
Manage skills and MCP servers for AI agents
Versions and profiles let you choose tools for a project, move configurations between environments and restore an earlier set.
Send a task to an external AI session with selected files and a budget limit
The context package records what the recipient received, so the result can be reviewed against the original brief.
A framework for organising knowledge used by AI agents
It helps agents find the right instructions, project context and code documentation, using links to sources to avoid conflicting copies.
Prepare CVs and applications from verified experience.
An AI agent receives selected sources and a task; documents are saved as separate revisions. Each vacancy is linked to the approved package and the exact text sent, so you can see what was sent and to whom.