Docs
Learn Nested
Nested runs specialized coding agents on Kubernetes you control. These docs cover the whole product: the concepts, every screen of the app, automations, and the platform underneath.
Start here
What Nested is, your first agent and task, and the guided tours.
- What is NestedNested runs specialized coding agents on Kubernetes you control. Learn what it is, who it is for, and how the app, console, and runtime fit together.
- QuickstartSign in, deploy your first agent, start a task, answer it from Needs you, and hand recurring work to an automation — in about ten minutes.
- Product toursFour short desktop walkthroughs — deploy an agent, connectors, scheduled tasks, workspaces and tasks — that you can start, skip, or replay anytime.
Concepts
The words Nested uses — agents, packs, workspaces, tasks, formulas, memories — and how they fit together.
- GlossaryThe words Nested uses — agent, pack, skill, formula, workspace, task, automation, connector, memory, team — and how they fit together.
- How tasks workA task’s life in Nested: running, needs you, completed, canceled, failed, or rejected — plus reopening, queued messages, subtasks, and automated tasks.
- TeamsEvery agent, workspace, and task belongs to a team. Create or join a team, invite members, set roles, and handle agents left behind by former members.
Chat and tasks
Workspaces, parallel task threads, formula runs, and the composer tools that make writing faster.
- Chat, workspaces, and tasksUse the task list, pick an agent and workspace in the composer, create folder or repository workspaces, and run several tasks in parallel.
- Needs you: tasks waiting on youWhen a task needs your answer, sign-in, or decision, Nested moves it to Needs you at the top of the task list — without pulling you out of your thread.
- Writing faster in chatRephrase a draft, dictate by voice, summarize a long reply, read it aloud, and turn any agent message into subtasks — without leaving the thread.
- Formula runs in chatA formula run opens on its own canvas: see where it is, act on checks, decisions, and inputs, open a step’s chat, and cancel or restart the run.
- The task panel, changes, and worktreesSee a task’s agent, session, memory, environments, and changed files beside the chat; review diffs; and run repository tasks in separate worktrees.
- Share, export, and runbooksShare a task with a public read-only link, export its history as Markdown, and read the runbook — the agent’s turn-by-turn log.
- Command palettePress Cmd/Ctrl+K to start a task with any agent, open an agent’s settings, jump to any page, or switch teams from the keyboard.
- The composerAttach files, reuse drafts, reference other tasks, set session options, run slash commands, choose memory, and queue messages while the agent works.
Your agents
Deploy an agent, then shape its settings, shared knowledge, memories, and machine.
- Deploy an agentDeploy a Nested agent: choose a base, packs, and a coding agent (Cursor, Claude Code, Kimi CLI), attach secrets, and choose who can chat with it.
- Agent settingsChange an agent you own: name, ACP agent, session defaults, access, packs, secrets, knowledge, runtime, automations, startup tasks, hooks, and activity.
- Shared knowledge centersAttach shared knowledge centers to an agent so it can read or update your docs, runbooks, or team wiki during tasks, with read-only or read-and-write access.
- MemoriesMemories hold what your agents learn about you and your work. Create, search, edit, restore, and reorganize them, and choose which tasks use them.
- Desktop and TerminalOpen an agent’s graphical desktop or a shell on its machine to watch work, verify results, or finish a sign-in, from the task panel or Settings → Agents.
- Agent sign-in, accounts, and auto-stopSign an agent in to its coding vendor, keep several vendor accounts and switch between them, and restart an agent that stopped after a quiet period.
- Startup tasks and lifecycle hooksRun messages every time an agent starts, and call shell scripts or HTTP endpoints before and after each session or prompt with lifecycle hooks.
- Site loginsSave website sign-ins your agents can use, including authenticator keys for two-step codes, and choose which agents may sign in without asking you.
- EnvironmentsSee the short-lived machines your agents create to run code, open their preview links, and keep, extend, or destroy them from a task or the Environments page.
- PacksBrowse your packs, your team’s packs, and the global catalog. Copy catalog packs, change visibility, review revisions, and attach packs to agents.
Automations
Run an agent on a schedule or when something happens — GitHub, Slack, webhooks, and more.
- AutomationsRun a Nested agent on a schedule or when something happens in GitHub, Slack, Sentry, or a webhook. Every run opens a task and lands in the automation’s history.
- Scheduled tasksHave a Nested agent start a prompt on chosen days and times or at a repeating interval, in a workspace you choose, with pause, run now, and history.
- WebhooksTurn inbound HTTP requests into Nested tasks: route, HMAC or token verification, Go-template task bodies with .JSON fields, deliveries, and redelivery.
- GitHubConnect the Nested GitHub App, then start agent tasks from pull requests, reviews, comments, pushes, and CI with automations or raw repository bindings.
- Slack botsConnect a Slack workspace to Nested, start agent tasks from mentions, channel messages, or emoji reactions, and post the agent’s answer back to Slack.
- SentryConnect a Sentry Internal Integration so new Sentry issues and events start Nested agent tasks, with signed webhooks and issue comments.
Your account
Overview, reports, preferences, secrets, notifications, connectors, and the pack builder.
- Overview of running workSee in-progress tasks, the latest weekly report, recent automation runs, token usage and estimated cost, and your agents on one Overview page.
- Task timelineBrowse a chronological, day-by-day history of tasks across all your agents and workspaces, with each task’s state, in the Nested task timeline.
- ReportsRead closed day, week, and month reports for your agents: task, automation, formula, and schedule counts, latency, missing days, and PNG or PDF downloads.
- Preferences and profileSet theme, text size, read-aloud, motion, and language in Nested, and manage your display name, phone, passkeys, and sign-out in your profile.
- Sign-in and passkeysSign in to Nested with a passkey, email and password, or Google; add and manage passkeys; and approve apps that ask to sign you in with Nested.
- SecretsStore API keys, tokens, and kubeconfigs as write-only secrets in Nested, keep them private or share them with your team, and attach them to agents.
- NotificationsGet notified when a task needs you, a task finishes, an agent’s status changes, or a formula needs a decision — in the browser or in the background.
- ConnectorsConnect the tools your Nested agents use on your behalf — OAuth or MCP connectors — reconnect expired ones, and add custom team connectors.
- Pack builderBuild a Nested pack on desktop: describe it, start from a template, import skills from Git, edit every section, verify the archive, and upload it to your team.
- DictationDictate messages in Nested with the browser’s speech service or Whisper running on your own device, and tune language, model, and your custom words.
- Activity logReview what you did in Nested: a filterable audit log of your account’s actions over about 90 days, with before-and-after detail for each change.
- Install the Nested appInstall Nested on your desktop or phone home screen as an app, enable background notifications on iOS, and reload when an update is available.
Platform
How Nested runs on your Kubernetes: architecture, the agent runtime, packs, security, and installation.
- ArchitectureWhat Nested installs into your Kubernetes cluster — control plane, agent pods, and platform services — and how a task flows through them.
- Agent runtimeHow a Nested agent pod works: coding agents over ACP, task sessions, the sandbox, idle stop and wake, crash recovery, and settings reloads.
- PacksThe pack format: skills, subagents, formulas, MCP servers, setup scripts, and guides in one verified, versioned archive you attach to agents.
- Formula referenceThe schema-2 formula format: typed input, steps, gates, checks, human input, state, limits, and how runs execute, fail, and restart.
- Tasks and agent collaborationHow Nested tasks work under the hood — states, claiming, runbooks — and how agents find each other and delegate work with A2A.
- Knowledge and memory internalsHow knowledge centers and user memories store pages, control access, keep versions, and how agents recall and save what they learn.
- MCP servers and connectorsWhere agent tools come from — the built-in Nested MCP server, connectors, and pack MCP servers — and how the catalog and Code Mode save tokens.
- Environments serviceHow Nested Environments create short-lived machines for agents: the manifest, Docker and Kubernetes runtimes, lifetimes, and isolation.
- Security modelHow Nested protects your data: deployment boundary, sign-in and tokens, encryption at rest, secrets, agent isolation, teams, and the audit log.
- Install NestedInstall a Nested cell on your Kubernetes with nested-install: requirements, the setup file, the install steps, upgrades, backups, and teardown.