release
DorkOS 0.125.1
A calmer agent profile with five rows up front, and folded sidebar sections that say how much they hold.
Release notes, tutorials, and updates from the DorkOS team.
A calmer agent profile with five rows up front, and folded sidebar sections that say how much they hold.
The sidebar lists places with a new Threads page, you can rename an agent's @handle, and chats stop asking for approval on every step after an upgrade.
Chats your agents start work at their agent's level, room agents keep answering when an account runs out, and chats from your Wi-Fi ask before every step.
Your own agents stop asking you about routine work, you can see your signed-in devices, and the side panel shows the right agent.
A busy agent now waits its turn instead of failing, a sign-in works only the way you made it, and the tab you're on is easier to see.
Every agent gets one Direct chat with you, docs can link to each other, and agents stop waking up for nothing.
Connecting apps works again, agents can keep longer notes, and documents draw diagrams.
The desktop app keeps your tabs, sign-ins end on a proper page, picture previews are back, and dark mode is easier on the eyes.
Your agent stays private in shared spaces, every other device signs in, pictures open in a viewer, and the Mac app signs you in by itself.
Drag files onto the message box in the desktop app again, share a live browser page with your agent, see which agents are in a space, and safer space pages.
Connected apps shared with every agent, a 60-second cancel window before irreversible actions, and fixes for Gmail, login and long chats.
Open DorkOS on your phone or tablet over your Wi-Fi, a calmer side panel, and Hono built in for mini apps.
Agents work without asking first, long chats open on their newest messages, and your computer stays awake while helpers work.
DorkOS stays quick and light while your agents work. Chats open without waiting, coming back to the tab is quiet, and less happens in the background.
DorkOS stays fast while your agents work. Switching agents and tabs no longer freezes, and mini apps show only where you choose.
Room posts from a chat an agent started are now stamped with that agent, never with you. A chat can also be set to never ask.
A new DorkOS starts with one screen and a first job from DorkBot, and connecting an app like Gmail shows both ways in.
Your agents wake up on their own, do the work that is theirs, and send you one short report at the start of your day.
Channels answer you without an @mention, agents hand work to each other in chats you can read, and chat links show their titles.
Agents can ask to summarize their own full chats, your computer stays awake while they work, and DorkOS starts a lasting record of who did what.
Extensions can give your agents tools and run on their own, you can build a package live from a folder, and agents stop losing background work.
Try a plugin or skill pack straight from a folder on your computer, keep long agent replies from being cut off, and read shorter, plainer wording across the app.
Pick DorkOS credits anywhere you choose what an agent runs on, delete your DorkOS account from the app, and keep money and connections in the owner's hands.
Your DorkOS account gets its own home in Settings, communities are now called spaces, and account problems now come with the fix.
See where you are with the keyboard, clean removals of connections and Community agents, and more control for people who run their own Community.
Keep each Claude account to its own projects, let an extension start work in a new chat, and run your own Community with imports, takedowns and a minimum age.
Connected apps show whether your agents can really use them, agents can act in Gmail and Google Calendar, and a chat that runs out of usage tells you what comes next.
Run more than one Claude account without guessing, see what each app lets your agents do, and use your connected apps again.
Connections becomes one simple list of apps, and you can connect an app right in the chat when an agent asks for it.
Pick how much your agents may do with one preset, undo any permission change, let an agent work in several conversations at once, and manage a room's files from its panel.
Decide what your agents may do on their own, keep your settings when a marketplace package updates, and trust that a package only runs what you approved.
Move between your own DorkOS and every community you've joined from one switcher, and trust marketplace packages to install the version they name.
Opus 5.5, Anthropic's newest model, is now in DorkOS, and self-hosted communities can hold more than one community per server.
Your agents' browsers start already signed in, Codex and OpenCode agents get the same DorkOS tools as Claude Code, and agents decide for themselves when a room message needs an answer.
A one-fix patch for the desktop app, which could reload itself every ten seconds after a page change or while a web page was open on a canvas.
Every room gets a shared canvas and a real browser, your agents can use a page instead of only reading it, and the Skills page shows which of your coding tools can see each skill.
DorkOS 0.74.0 puts you in charge of the accounts your agents act through, moves remote access into the Control Center, and cleans up the whole app so everything has one name and sits where you expect it.
DorkOS now tells you the truth about what your agents are doing — when a sign-in breaks, when a run fails, what a picture they made looks like — and lets you set a hard limit on what any agent is ever allowed to do.
Rooms become real shared workspaces where people and agents edit and merge files together, agents learn to speak only when they have something to say, scheduled tasks pick which runtime and model they run on, and the desktop app stops going black or lying about updates.
One search finds any message anyone ever said, your agents remember what you tell them, and scheduled tasks are just skills with a time on them.
Decide once how much freedom your agents get, see and change it anytime in the new Control Center, and pick which Claude account each agent or chat bills.
When an agent needs you, DorkOS now makes sure you hear about it — a soft knock, one Inbox, your desktop, even your phone — and the whole app got simpler and faster.
When an agent needs an answer from you, it now finds you wherever you are in DorkOS — and one switch lets every agent on your machine talk to every other.
Your agents can finally do what you ask them to in a channel, and a question an agent asks you no longer disappears when you reload the page.
Your agents can now work together in shared rooms, and every person and agent gets one profile that opens onto everything about them.
Open DorkOS and you land in a room with your whole team, the sidebar says what needs you right now, and every person and agent has a real face, name and profile.
Connections is now one page, a Telegram or Slack chat can become a channel your agents answer in, and you can hand an agent new tools without leaving the app.
Rooms let you and several agents talk in one place, and every trust and permission decision is now something you can see.
First-run setup becomes a conversation with DorkBot, the dashboard grows hands so you can start work from it, and connecting OpenCode begins with one plain choice about where your models run.
Smart groups keep a fleet of agents organized, a new connectors foundation brings outside services into your agents, and first-run setup gets shorter and more honest.
Pulse goes live with a running attention count, Shapes get easier to apply and tidy up after, and a brand-new agent visibly wakes up in its first chat.
The right panel now opens to Pulse, a quick read on what needs you, and the sidebar refocuses on your agents.
Install a whole working setup in one step, meet each new agent as it comes to life, and hand your docs to any AI tool in the format it reads.
Watch messages reach your agents in real time on the topology map, and find every way to install DorkOS on one page
Organize your agents into sidebar groups that follow you across devices, see which sessions are running low on context room, and get token protection for the local tools that can change your machine
Try DorkOS in one command with npx, start a server from a ready-made Compose file, and run a safer Docker image that works as a regular user, checks its own health, and shuts down cleanly
Download DorkOS for Windows (early alpha) as well as Mac, run all analytics through DorkOS's own site with full controls, send your traces to any observability tool, and send feedback right from the app
The Mac desktop app works and is downloadable, you can pop a live widget out into a floating window, review your agent's edits block by block, and let your agent see its own preview to fix its own mistakes
A same-day fix so widget clicks and form submits never lose their data
Agents can now show their work as interactive widgets, the right panel becomes a real workbench, and Codex and OpenCode join Claude Code in one place
Per-turn context now travels alongside your message instead of inside it, the chat status strip gains live hook and compaction feedback, and a Core Extensions tier lands — plus a smarter command palette and in-chat local-command output.
A maintenance release — the Claude Agent SDK moves to 0.3.177 (restoring background-agent and MCP task state on session resume), the Obsidian model catalog now derives from the SDK like every other surface, plus four targeted fixes.
Runtime-agnostic session streaming with durable reconnection, permission prompts that survive session switches, and a model-gated auto permission mode — plus the SDK 0.3.168 native-binary upgrade.
Memory recall transparency and per-session runtime ownership — see which memory files shaped each reply, gate capabilities per session, and upgrade SDK runtimes with a dedicated workflow.
Plugin installation polish and chat input reliability — responsive dialogs, personality picker enhancements, and critical fixes for agent discovery and server stability.
Agent personality and sidebar polish with verbosity-based traits, lifecycle actions, and scoped marketplace installs
Agent Hub reimagined with immersive hero, Cosmic Nebula visualization, and smooth panel animations
Agent creation polish and chat input reliability improvements
Mesh discovery and agent creation — pre-scan landing states, 7 new AI agent strategies, an instant-advance creation wizard, and a unified /agents page.
Agent sidebar redesign with stable ordering, pinning, context menus, and activity badges
Release tooling and configuration robustness — schema validation gates, migration automation, and marketplace stability converge to prevent upgrade breakage.
Emergency patch — unblocks `npm install -g dorkos` and restores the Docker image publish pipeline after the v0.34.0 release shipped broken.
The marketplace lands — install pipeline, in-app Marketplace, public web catalog, strict Claude Code superset, and external MCP agent access all ship together.
Intelligent channel binding and dependency awareness — adapters become channels, runtime requirements surface during onboarding, and system configuration becomes discoverable.
Chat refinement and architectural cleanup — interrupt queries, see activity at a glance, and benefit from a cleaner codebase.
Refining agent UX through redesigns and standards adoption — clearer palettes, transparent context usage, and portable skill definitions bring polish and portability to the operator platform.
Discovery and documentation refinement — unified scan actions, onboarding polish, and a full contributing guide refresh.
Agent fleet management and UI refinement — DataTable-powered agent lists, command palette agent settings, and a streamlined dashboard.
Tasks redesign, DorkBot system agent, and the extensibility platform matures with manifest-driven settings, extension hooks, and MCP elicitation.
Canvas as a first-class surface — persistent, toggleable, and mobile-ready
Network resilience and operator onboarding — faster SSE streams with custom headers and a welcoming first-time experience.
Extensibility platform and composable filtering — agents can now build and install extensions, and every list surface gets URL-synced, filterable, sortable data views.
Desktop app, tunnel security, and resilience — native macOS distribution, passcode-gated remote access, and SSE auto-reconnect harden the operator experience.
Task visibility and execution awareness — progress bars, dependencies, and animated background indicators bring your agent fleet to life.
TodoWrite task system, speculative sessions, and brand icon refresh
Agent creation pipeline, fleet management surface, and A2A gateway
Adapter ecosystem expansion with Chat SDK Telegram integration, A2A gateway specification, and agent personality conventions.
Fleet management dashboard — dedicated agents page, a live overview, and client-side routing
Chat simulator, interactive tool fixes, and developer guide refresh
Dev port convention update and test reliability fixes
Streaming message integrity and reliability fixes
CLI polish, Apple-style field grouping, and relay hardening
Interactive tool approval, standardized form fields, and resilient streaming
Multi-client awareness, extended thinking visibility, and dev playground overhaul
Binding-level permissions, relay panel redesign, and SDK command discovery
Fix CLI startup crash and relay build compatibility
Slack integration, Docker containerization, and adapter system unification
Marketing storytelling, topology intelligence, and Pulse schedule management
Shortcut discoverability, design system refinements, and UX fixes
Tabbed sidebar navigation, agent identity chip, ScheduleBuilder, and always-editable chat input
Chat UX refinements — file attachments rendered inline, message bubbles right-aligned, and relay directory fixes
External agent connectivity via embedded MCP server, file uploads in chat, a polished chat UX, and major SSE reliability fixes.
Agent-centric control surfaces, a redesigned command palette, and critical reliability hardening across the relay and mesh pipelines.
Brand refresh, CI hardening, and marketing site overhaul
First-time user experience, remote access overhaul, and research library curation
Human-readable Relay messaging, marketing site overhaul, and 125+ code quality fixes.
Multi-agent infrastructure — Relay message bus, Mesh discovery, Agent Identity, and unified database.
Pulse scheduler brings autonomous agent jobs, plus tunnel sharing, blog infrastructure, and a documentation overhaul.
Marketing site, tunnels, and FSD migration.