{"id":47261898,"url":"https://github.com/sapienxai/agentos","last_synced_at":"2026-04-02T12:00:33.455Z","repository":{"id":342699419,"uuid":"1174794422","full_name":"SapienXai/AgentOS","owner":"SapienXai","description":"Human Control Layer for AI Agents. Built on OpenClaw by SapienX","archived":false,"fork":false,"pushed_at":"2026-03-24T09:30:05.000Z","size":3409,"stargazers_count":6,"open_issues_count":0,"forks_count":1,"subscribers_count":1,"default_branch":"main","last_synced_at":"2026-03-24T21:32:54.564Z","etag":null,"topics":["agent","agentos","agents","ai","ai-agent","ai-agents","ai-assistant","ai-tools","mission","mission-control","openclaw","openclaw-plugin","openclaw-skill","openclaw-skills"],"latest_commit_sha":null,"homepage":"https://sapienx.app/agentos/","language":"TypeScript","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/SapienXai.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2026-03-06T21:10:32.000Z","updated_at":"2026-03-24T09:30:10.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/SapienXai/AgentOS","commit_stats":null,"previous_names":["sapienxai/agentos-core","sapienxai/agentos-missioncontrol","sapienxai/agentos"],"tags_count":9,"template":false,"template_full_name":null,"purl":"pkg:github/SapienXai/AgentOS","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SapienXai%2FAgentOS","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SapienXai%2FAgentOS/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SapienXai%2FAgentOS/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SapienXai%2FAgentOS/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SapienXai","download_url":"https://codeload.github.com/SapienXai/AgentOS/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SapienXai%2FAgentOS/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31305971,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-02T09:48:21.550Z","status":"ssl_error","status_checked_at":"2026-04-02T09:48:19.196Z","response_time":89,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent","agentos","agents","ai","ai-agent","ai-agents","ai-assistant","ai-tools","mission","mission-control","openclaw","openclaw-plugin","openclaw-skill","openclaw-skills"],"created_at":"2026-03-14T23:45:18.266Z","updated_at":"2026-04-02T12:00:33.449Z","avatar_url":"https://github.com/SapienXai.png","language":"TypeScript","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"public/readme/banner.jpeg\" alt=\"AgentOS mission-control interface\" width=\"100%\" /\u003e\n\n# AgentOS | Mission Control\n\n**Human operating layer for coordinating AI agents, projects, and companies from a single workspace.**\n\nBuilt on top of OpenClaw, the agent orchestration kernel.\n\n\u003cp\u003e\n  \u003ca href=\"https://sapienx.app/agentos\"\u003e\u003cstrong\u003eWebsite\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"https://www.youtube.com/watch?v=ujz-4bYDjdY\"\u003e\u003cstrong\u003eWatch Demo\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#why-agentos\"\u003e\u003cstrong\u003eWhy AgentOS\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#quick-start\"\u003e\u003cstrong\u003eQuick Start\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#architecture\"\u003e\u003cstrong\u003eArchitecture\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#key-features\"\u003e\u003cstrong\u003eFeatures\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#product-highlights\"\u003e\u003cstrong\u003eHighlights\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#setup-and-development\"\u003e\u003cstrong\u003eSetup\u003c/strong\u003e\u003c/a\u003e\n  ·\n  \u003ca href=\"#roadmap\"\u003e\u003cstrong\u003eRoadmap\u003c/strong\u003e\u003c/a\u003e\n\u003c/p\u003e\n\n\u003cp\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Next.js-16-0b1220?style=for-the-badge\u0026logo=nextdotjs\u0026logoColor=white\" alt=\"Next.js 16\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/React-19-07111d?style=for-the-badge\u0026logo=react\u0026logoColor=61dafb\" alt=\"React 19\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/TypeScript-Strict-0f172a?style=for-the-badge\u0026logo=typescript\u0026logoColor=3178c6\" alt=\"TypeScript\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/OpenClaw-Kernel-111827?style=for-the-badge\" alt=\"OpenClaw kernel\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Local--First-Control_Plane-101828?style=for-the-badge\" alt=\"Local-first control plane\" /\u003e\n\u003c/p\u003e\n\n\u003c/div\u003e\n\n## Why AgentOS\n\nAs AI agents become cheaper to run, the bottleneck shifts from raw orchestration to human control.\nSomeone still has to decide what matters, inspect active work, route missions, review outputs, and keep multiple projects legible.\n\nMost agent systems expose runtimes, sessions, and CLI primitives.\nAgentOS adds the missing operating layer above them: a mission-control interface for humans coordinating teams of agents across real workspaces.\n\nThis repository contains the current AgentOS control plane: a Next.js application that sits above OpenClaw and turns live agent state into an operator-facing system for planning, execution, inspection, and workspace management.\n\n## The Problem It Solves\n\nRunning one agent is not the hard part.\nOperating many agents across many projects is.\n\nAgentOS is built for that coordination problem:\n\n- A human operator needs one place to see workspaces, agents, models, runtimes, and health.\n- Missions should map to real project folders, not ephemeral chat threads.\n- Runtime output should be inspectable after the fact, including created files and transcript history.\n- Agent teams need structure: presets, policies, memory, workspace scaffolds, and repeatable operating conventions.\n- As the \"one-person company\" model emerges, the human needs a control layer, not just an orchestration engine.\n\n## Quick Start\n\nInstall the packaged launcher:\n\n```bash\npnpm add -g @sapienx/agentos\nagentos start --open\nagentos doctor\n```\n\nRun the app locally from this repository:\n\n```bash\npnpm install\npnpm dev\n```\n\nIf OpenClaw is not ready yet, AgentOS starts in an explicit onboarding or fallback path instead of pretending a live control plane exists.\n\n## Architecture\n\n```mermaid\nflowchart TD\n    Human[\"Human Operator\"] --\u003e AgentOS[\"AgentOS\u003cbr/\u003econtrol layer / operating layer\"]\n    AgentOS --\u003e OpenClaw[\"OpenClaw\u003cbr/\u003eagent orchestration kernel\"]\n    OpenClaw --\u003e Runtime[\"LLMs, tools, channels, automations, agents\"]\n```\n\n### Layer Responsibilities\n\n| Layer | Responsibility |\n| --- | --- |\n| Human operator | Sets direction, reviews work, approves risky actions, and steers the system |\n| AgentOS | Presents topology, planning, inspection, workspace bootstrap, settings, and mission dispatch |\n| OpenClaw | Owns agent orchestration, gateway state, models, sessions, channels, and execution surfaces |\n| LLMs and tools | Perform the underlying reasoning and tool-backed work |\n\n### Control Plane Shape\n\n```mermaid\nflowchart LR\n    UI[\"AgentOS UI\u003cbr/\u003eSidebar / Canvas / Inspector / Command Bar / Planner\"] --\u003e API[\"Next.js App Router + API routes\"]\n    API --\u003e SERVICE[\"OpenClaw service adapter\u003cbr/\u003esnapshot normalization + write actions\"]\n    SERVICE --\u003e CLI[\"OpenClaw CLI\"]\n    CLI --\u003e GATEWAY[\"Gateway status + presence\"]\n    CLI --\u003e CONFIG[\"Agent config + workspace bindings\"]\n    CLI --\u003e SESSIONS[\"Sessions + transcript files\"]\n    SERVICE --\u003e FS[\"Workspace filesystem + .mission-control state\"]\n    API --\u003e STREAM[\"SSE snapshot stream\"]\n    STREAM --\u003e UI\n```\n\n## AgentOS and OpenClaw\n\nOpenClaw is the kernel.\nIt handles the underlying agent runtime, CLI, gateway, models, sessions, automations, and execution primitives.\n\nAgentOS is the operating layer above it.\nIt does not replace OpenClaw.\nInstead, it reads live OpenClaw state, normalizes it into a control-plane snapshot, and gives the human operator a coherent surface for acting on that state.\n\nIn practice, that means:\n\n- OpenClaw remains the source of truth for agents, sessions, models, and gateway status.\n- AgentOS translates UI actions into real OpenClaw commands and real filesystem changes.\n- AgentOS is intentionally not a mock dashboard; it is a control surface over live operational state.\n\n## How The System Works\n\n1. AgentOS reads live OpenClaw surfaces such as gateway status, agent inventory, config, models, sessions, presence, and transcript files.\n2. The service layer normalizes that data into a single `MissionControlSnapshot`.\n3. The UI renders that snapshot as a mission-control surface with a topology canvas, sidebar, inspector, and command bar.\n4. Operator actions such as mission dispatch, workspace creation, agent updates, planner deploys, gateway changes, or file reveal calls are translated into OpenClaw CLI commands and local filesystem operations.\n5. Snapshot state is refreshed over Server-Sent Events so the UI can stay close to real runtime activity.\n\n## Key Features\n\n- Live topology canvas for real workspace -\u003e agent -\u003e runtime relationships.\n- Mission dispatch that targets real OpenClaw agents and supports thinking levels.\n- Transcript-backed runtime inspection, including final output, warnings, token usage, and created files.\n- File reveal actions from the inspector for artifacts written to the local filesystem.\n- Workspace wizard with basic create flow and advanced planner mode, including source modes (`empty`, `clone`, `existing`), templates, team presets, model profiles, and kickoff missions.\n- Structured workspace scaffolding with `AGENTS.md`, `SOUL.md`, `IDENTITY.md`, `TOOLS.md`, `HEARTBEAT.md`, `MEMORY.md`, `docs/`, `memory/`, `deliverables/`, `skills/`, and `.openclaw/project-shell/`.\n- Agent creation and editing with policy presets (`worker`, `setup`, `browser`, `monitoring`, `custom`) plus heartbeat, file-access, install-scope, and network controls.\n- Guided workspace planner that models company, product, workspace, team, operations, and deploy decisions inside the workspace wizard.\n- Planner deploy flows that can turn a plan into a live workspace, agent team, automations, channels, and first missions.\n- OpenClaw onboarding, model setup, gateway control, reset, and update flows directly from the UI.\n- Configurable gateway endpoint and default workspace root from settings.\n- Explicit fallback mode when OpenClaw is unavailable, rather than pretending live control exists.\n\n## Product Highlights\n\nThree flows define the current AgentOS experience:\n\n### One-Click OpenClaw Setup\n\n\u003cimg src=\"public/readme/setup.webp\" alt=\"Guided OpenClaw setup and onboarding flow\" width=\"100%\" /\u003e\n\nGo from zero to a live control plane in minutes. AgentOS detects what is missing, installs OpenClaw, and guides operators through system and model onboarding without the usual setup friction.\n\n### AI Workspace Architect\n\n\u003cimg src=\"public/readme/create.webp\" alt=\"AI architect flow for creating workspaces, tasks, and agents\" width=\"100%\" /\u003e\n\nTurn a rough idea into an operational blueprint. The architect flow uses OpenClaw-backed planning to shape workspaces, tasks, and specialized agent roles before execution even begins.\n\n### Guided Workspace Wizards\n\n\u003cimg src=\"public/readme/wizzard.webp\" alt=\"Workspace wizard for project context and agent setup\" width=\"100%\" /\u003e\n\nLaunch new projects with confidence. Structured wizards capture context, scaffold the right workspace shape, and assemble the agent team your project needs without manual busywork.\n\n## UI Surfaces\n\n| Surface | Purpose |\n| --- | --- |\n| `MissionSidebar` | Gateway diagnostics, workspace navigation, models, agents, and workspace or agent CRUD |\n| `MissionCanvas` | Visual topology for workspaces, agents, and runtimes with selection and mission feedback |\n| `InspectorPanel` | Detailed inspection of selected entities, transcript output, raw payloads, and created files |\n| `CommandBar` | Mission composition, agent targeting, thinking level selection, refresh, and quick suggestions |\n| `WorkspaceWizardDialog` | Handle both basic workspace creation and advanced planner-driven workspace design and deploy |\n| `OpenClawOnboarding` | Detect, install, start, verify OpenClaw, and guide model readiness when the local machine is not ready |\n| `ResetDialog` | Preview Mission Control reset or full uninstall actions and stream execution progress and logs |\n\n## Repository Map (Key Files)\n\n```text\napp/\n  api/\n    agents/\n    diagnostics/\n    files/reveal/\n    gateway/control/\n    mission/\n    onboarding/\n    onboarding/models/\n    planner/\n    reset/\n    runtimes/[runtimeId]/\n    settings/\n    snapshot/\n    stream/\n    system/open-terminal/\n    update/\n    workspaces/\n  layout.tsx\n  page.tsx\n\ncomponents/mission-control/\n  canvas.tsx\n  command-bar.tsx\n  create-agent-dialog.tsx\n  inspector-panel.tsx\n  mission-control-shell.tsx\n  openclaw-onboarding.tsx\n  operation-progress.tsx\n  reset-dialog.tsx\n  sidebar.tsx\n  nodes/\n  workspace-wizard/\n    workspace-wizard-dialog.tsx\n    workspace-wizard-draft-pane.tsx\n    workspace-wizard-header.tsx\n    wizard-composer.tsx\n    wizard-message-list.tsx\n    wizard-suggestion-chips.tsx\n\nhooks/\n  use-mission-control-data.ts\n  use-workspace-wizard-draft.ts\n\nlib/openclaw/\n  agent-heartbeat.ts\n  cli.ts\n  agent-presets.ts\n  fallback.ts\n  operation-progress.ts\n  planner.ts\n  planner-core.ts\n  planner-presenters.ts\n  presenters.ts\n  readiness.ts\n  reset.ts\n  service.ts\n  types.ts\n  workspace-presets.ts\n  workspace-wizard-inference.ts\n  workspace-wizard-mappers.ts\n\npackages/agentos/\n  bin/\n  scripts/\n  README.md\n  package.json\n```\n\nThis is a representative map of the current control-plane code, not an exhaustive file listing.\nMany internal files still use `mission-control` naming; that is the current AgentOS application shipped in this repository.\n\n## Setup And Development\n\n### Prerequisites\n\n- A recent Node.js runtime\n- `pnpm`\n- OpenClaw installed locally and reachable on `PATH`\n\nIf OpenClaw is installed in a non-standard location:\n\n```bash\nexport OPENCLAW_BIN=/absolute/path/to/openclaw\n```\n\n### Install\n\nGitHub Release installer:\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/SapienXai/AgentOS/main/install.sh | bash\nagentos start --open\nagentos stop\nagentos doctor\n```\n\nInstall a specific published version:\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/SapienXai/AgentOS/main/install.sh | AGENTOS_VERSION=0.1.9 bash\n```\n\nPackage manager install:\n\n```bash\npnpm add -g @sapienx/agentos\n# or\nnpm install -g @sapienx/agentos\n\nagentos start --open\nagentos stop\nagentos doctor\n```\n\nStop a running server:\n\n```bash\nagentos stop\n```\n\nUninstall:\n\n```bash\nagentos uninstall\n```\n\nIf AgentOS was installed with `pnpm` or `npm`, remove it with your package manager instead:\n\n```bash\npnpm remove -g @sapienx/agentos\n# or\nnpm uninstall -g @sapienx/agentos\n```\n\nLocal development:\n\n```bash\npnpm install\nopenclaw --version\nopenclaw gateway status --json\n```\n\nIf the gateway service is missing or not loaded:\n\n```bash\nopenclaw gateway install --json\nopenclaw gateway status --json\n```\n\n### Releases\n\nPush a tag in the format below to build platform-specific release assets on GitHub Releases:\n\n```bash\ngit tag agentos-v0.1.9\ngit push origin agentos-v0.1.9\n```\n\nThe release workflow uploads:\n\n- `agentos-darwin-arm64.tgz`\n- `agentos-darwin-x64.tgz`\n- `agentos-linux-x64.tgz`\n- matching `.sha256` files\n\n### Run The App\n\n```bash\npnpm dev\n```\n\nOpen the URL printed by Next.js, typically:\n\n```text\nhttp://localhost:3000\n```\n\nIf OpenClaw is unavailable when the app starts, AgentOS can fall back to a demo snapshot and exposes an in-app onboarding flow to help bring the local machine online.\n\n### Quality Checks\n\n```bash\npnpm lint\npnpm typecheck\npnpm build\n```\n\n## Operational Notes\n\n- AgentOS is currently local-first. Several API routes spawn local processes, inspect transcript files, and mutate workspace directories.\n- This makes the current implementation best suited for operator workstations or trusted environments, not serverless-only deployments.\n- OpenClaw remains the primary runtime source of truth; AgentOS adds control-plane state rather than a separate database layer.\n- The app is configured for standalone Next.js output via `next.config.mjs`.\n\n## Control-Plane APIs\n\n| Route | Method | Purpose |\n| --- | --- | --- |\n| `/api/snapshot` | `GET` | Return the normalized AgentOS snapshot |\n| `/api/stream` | `GET` | Stream snapshot updates over SSE |\n| `/api/diagnostics` | `GET` | Return gateway diagnostics and presence |\n| `/api/mission` | `POST` | Dispatch a mission to a real OpenClaw agent |\n| `/api/agents` | `GET`, `POST`, `PATCH`, `DELETE` | Read and mutate agents |\n| `/api/workspaces` | `GET`, `POST`, `PATCH`, `DELETE` | Read and mutate workspace projects |\n| `/api/runtimes/:runtimeId` | `GET` | Load transcript-backed runtime output |\n| `/api/onboarding` | `POST` | Install or start OpenClaw and verify readiness |\n| `/api/onboarding/models` | `POST` | Discover models, refresh readiness, set a default model, or guide provider login |\n| `/api/update` | `POST` | Run `openclaw update` and stream output |\n| `/api/gateway/control` | `POST` | Start, stop, or restart the OpenClaw gateway |\n| `/api/planner` | `POST` | Create a new workspace planning draft |\n| `/api/planner/:planId` | `GET`, `PUT` | Load or save a planning draft |\n| `/api/planner/:planId/turn` | `POST` | Process a planner conversation turn |\n| `/api/planner/:planId/simulate` | `POST` | Simulate the planner team |\n| `/api/planner/:planId/deploy` | `POST` | Deploy a planned workspace |\n| `/api/reset` | `POST` | Preview or execute a Mission Control reset or full uninstall flow |\n| `/api/settings/gateway` | `PATCH` | Update the OpenClaw gateway endpoint |\n| `/api/settings/workspace-root` | `PATCH` | Update the default workspace root |\n| `/api/system/open-terminal` | `POST` | Open a supported OpenClaw command in Terminal on macOS |\n| `/api/files/reveal` | `POST` | Reveal a local file in Finder, Explorer, or the platform file manager |\n\n## Local State And Persistence\n\nAgentOS keeps most durable operational state close to the workspace and to OpenClaw itself.\n\n- OpenClaw-backed runtime state comes from gateway status, agent config, models, sessions, presence, and transcript files.\n- AgentOS settings are stored in `.mission-control/settings.json`.\n- Planner drafts and planner runtime assets are stored under `.mission-control/planner/`.\n- Planner deploys write workspace-specific planning artifacts under `.openclaw/planner/`, including `blueprint.json` and `deploy-report.json`.\n- Browser convenience state such as theme, draft missions, recent prompts, and the last planner id is stored in `localStorage`.\n- When OpenClaw is unavailable, AgentOS returns an explicit fallback snapshot with demo workspaces, agents, models, and runtimes.\n\n## Screens And Workflows Worth Exploring\n\n- Create a workspace from scratch and inspect the generated scaffold files.\n- Open the workspace wizard in advanced mode and move from company context to deploy.\n- Create agents with different presets and heartbeat policies.\n- Dispatch a mission, then inspect runtime output and created files from the inspector.\n- Change the gateway endpoint or workspace root from settings and watch the live snapshot refresh.\n\n## Roadmap\n\nThis repository already shows the shape of a broader operating system for AI work.\nDirectionally, the next layer looks like this:\n\n- Deeper company-level operations above individual project workspaces.\n- Richer provisioning for channels, automations, hooks, and recurring operational loops.\n- Stronger governance, permissions, approvals, and audit trails for multi-agent work.\n- Better remote and multi-host control over OpenClaw-backed environments.\n- More durable historical views for runtime analytics, operational memory, and handoff quality.\n\n## Contributing\n\nContributions are welcome.\nIf you want to extend the control plane, the planner, the workspace bootstrap flow, or the OpenClaw integration, open an issue or pull request.\n\nPlease keep contributions aligned with the current design principles:\n\n- Keep the project developer-focused and operationally grounded.\n- Prefer real OpenClaw-backed behavior over front-end-only mocks.\n- Keep user-facing copy and documentation in English.\n- Run `pnpm lint`, `pnpm typecheck`, and `pnpm build` before opening a PR.\n- Prefer concise English commit messages; Conventional Commits are a good fit here.\n\n## License\n\nThis repository does not currently include a license file.\nUntil one is added, assume standard copyright restrictions apply.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsapienxai%2Fagentos","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsapienxai%2Fagentos","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsapienxai%2Fagentos/lists"}