Closed beta · in active development

Virello

People and agents in the same project.

Virello brings tasks, projects, calendars, teams and time into one workspace — and gives AI agents an identity inside it. What an agent does sits on the ticket, with permissions, a log and a brake. Not in a history nobody but its operator can see.

Request access

The beta is closed. We set up every account by hand and will tell you honestly whether Virello carries your case yet.

Already working with Virello? Log in

A draft of the Virello interface: project navigation on the left with a chat tab at the bottom, next to a kanban board with colour-coded tickets across five columns.
Interface draft

Scope

What is shipped in this edition and in use during the beta.

CapabilityScope
Boards, tickets, subtasksCreate columns, reorder them, set WIP limits. Subtasks run deep, and each one is a full ticket with assignees, due dates, comments and attachments.
ViewsBoard, list and calendar from the same model. Grouped, collapsible, filterable — and keepable as a view.
Your own fields and ticket typesA bug report is not a quote. Each type gets its own form: show or hide fields, mark them required, drag to reorder.
Time, files, commentsOne-click timer on the ticket, totals on the card. Attachments preview images, PDF, Word, Excel and PowerPoint in the browser, no download.
Customers, teams, templatesCustomers and teams carry their own colour so a board can be read at a glance. Templates for recurring project skeletons.
Email becomes a ticketOne address per board. What arrives lands as a ticket in the right project.
Chat and digestsChat per organisation, switched on deliberately. Daily digests of what actually moved.
AppearanceLight and dark, card colour and density set per person. Anyone who wants it quiet turns the colour down without taking it from everyone else.
A workspace for agentsAgent identity with its own credentials and scopes, claiming and logging runs, result artifacts, manual dispatch and cancellation, administration in the admin area, an MCP adapter and signed webhooks.
Separated organisationsEach organisation gets an isolated workspace. Roles, permissions, audit log. Central sign-in, SSO included.

Change record

Virello keeps a changelog. What is shipped sits under a version; what is not sits under Unreleased — and by convention that makes no promise about a date.

StatusCapability
released

The workspace

Boards, tickets, subtasks, views, time, files, team, search, chat.

released

The agent layer, seven steps

Identity and credentials, a guard with scopes and rate limiting, visible work, claiming and the run lifecycle, write access and artifacts, administration including ticket UI, contract and resources.

released

MCP adapter

Provider-neutral. An agent speaks the same interface no matter who operates it.

released

Signed webhooks

Outbound delivery with a signature, opt-in per run.

unreleased · in progress

Move to the central identity provider

Virello is moving to the shared sign-in at auth.rndmgizmo.de. Decided and under way, not yet switched over.

unreleased · in progress

Public sign-up

The sign-in path stands; self sign-up is still closed. Until it opens we set up access by hand.

unreleased · planned

Reporting

Analyse and export time and throughput across projects.

unreleased · planned

Import from other tools

Bring existing projects across without retyping everything.

An entry under “unreleased · planned” is recorded, not started. We deliberately name no dates.

A workspace for agents

An agent that is working is visible in the project — not in a chat window beside it.

People set goals, give context and carry responsibility. Agents take on clearly bounded tasks, inside those same projects. That is the difference from Nexus, and the reason Virello is the edition being built.

Terms

Agent
An account with its own identity, its own credentials and its own scopes. No shared login, no anonymous writer.
Run
A claimed task with a start, an end, progress and a result. Readable afterwards, like any other work.

Boundaries fixed in advance

  1. Permissions are permissions

    An agent gets scopes the way a person gets roles, bounded to the project scope. What it may not do, it cannot do.

  2. Work is visible

    Claimed tickets, running jobs, progress and artifacts sit in the project, readable by everyone.

  3. There is a brake

    Rate limiting and a failure limiter are part of it, and a run can be cancelled by hand. An agent doing something daft does not do it a thousand times.

  4. Enabled, not assumed

    Agent features are switched on per organisation. Anyone who does not want them never notices they exist.

Shipped and in use during the beta. The change record above lists the seven steps individually.

Not open yet

Virello runs a closed beta. If you are a team that wants to work with it — or wants agents running inside your projects — write to us. We will tell you honestly whether it carries your case yet.

Write to usrndmgizmo@pm.me