Project / Rodon Workspace

One place for intelligence to become work.

Rodon is organized as a workspace rather than a collection of AI features. Conversations begin the work; documents, tasks, knowledge, media, and operations allow it to continue.

01

Conversations

Choose the right model, agent, context, and tools without rebuilding the working environment every time.

02

Durable work

Turn useful output into documents, notes, tasks, events, and source-linked records.

03

Background operations

Follow research, transcription, computation, and agent actions after the initiating window is closed.

Applications / Capabilities

A workspace, not another dashboard.

The applications follow the lifecycle of work. Creation tools sit beside knowledge and action tools, while operational state remains visible without dominating the interface.

  1. Create and think through Chat, Documents, Sound Studio, and Media.
  2. Organize durable intent through Tasks, Calendar, Mail, and Library.
  3. Extend capability through Agents, Marketplace, Models, MCP, Compass, and Compute.
  4. Inspect ongoing work through Operations instead of scattering progress indicators across applications.

Selected capabilities

Agents

Agents with a defined scope.

Agents combine selected models, instructions, knowledge, skills, and tools for a clear purpose.

  • Tools and context are assigned intentionally.
  • Important actions can require approval and remain visible in Operations.

Documents / Durable work

Useful thought should outlive the conversation.

Documents provide an editable, versioned destination for work that comes out of conversations or research.

  • Sources and originating conversations can remain attached.
  • Changes can be reviewed, restored, exported, and indexed.

Compute

Execution in a separate environment.

Rodon Compute gives model-generated analysis an isolated environment for code, data, files, and artifacts.

  • Sessions begin without application secrets or trusted host access.
  • Access and resources are policy-bound, while runs remain visible.

Architecture / Shared foundations

Different capabilities. One shared language.

04ApplicationsConversations · Documents · Notes · Tasks · Calendar · Mail · Media
03Intelligence and actionModels · Agents · Skills · MCP · Compute · Approved actions
02Knowledge and operationsLibrary · Search · Memory · Sources · Synchronization · Background work
01Trust and durable objectsIdentity · Ownership · Permissions · Provenance · Audit · Storage

Shared identity, provenance, permissions, search, Operations, and audit concepts keep every application connected without forcing it into the same interface.

Information flow

Context moves with the work.

Information can change form without losing why it exists, who owns it, or where it came from.

  • Mail can become a task while retaining its source.
  • Research can become a source-backed document.
  • A recording can become a timestamped Library transcript.
  • An approved task can become a visible agent action.

Principles

Clear boundaries, visible behavior.

01

Your context remains yours.

Provider choice, local infrastructure, permissions, storage, import, and export are product capabilities rather than footnotes.

02

Actions should be visible.

Important agent work requires clear intent, approval, cancellation, source attribution, and understandable audit history.

03

Memory should be inspectable.

People can see what was remembered, where it came from, when it was used, and whether it should remain available.

04

Tools should share one language.

Documents, notes, tasks, events, files, sources, operations, and actions use shared ownership and provenance concepts.

05

Execution should be bounded.

Generated code belongs in isolated compute. Trusted development infrastructure is never a silent fallback.

06

Degraded states should be honest.

Missing capability, uncertain cancellation, synchronization errors, and unavailable providers are shown rather than disguised.

Continue

Technical reference

Documentation →

Technical boundaries, APIs, integrations, protocols, and deployment guidance.

Security →

Concrete trust boundaries, approvals, isolation, and deployment considerations.

Open workspace →

Continue to the configured Rodon Workspace application.