14 packages on npm — create an AI team in one API call

We handle the
boring stuff.
You change the world.

Agent execution, workflows, connectors, memory, auth, budgets — all taken care of.
Focus on the exciting problems only you can solve.

Get Started
$ npx create-boringos my-app
index.ts
// Create an AI workforce in 5 lines

const app = new BoringOS({});
await app.listen(3000);

// Spin up a full engineering team — CEO delegates to CTO,
// CTO assigns to engineers, QA validates. All autonomous.
await createTeam(db, "engineering", { tenantId });

// Give them a goal
await createTask(db, {
  title: "Build the MVP",
  description: "Ship user auth, dashboard, and billing by Friday.",
  assigneeAgentId: ctoId,  // CTO breaks it down, delegates to engineers
});

// CEO wakes up. Reads the task. Delegates to CTO.
// CTO creates subtasks. Assigns to engineers.
// Engineers write code. QA reviews. Tasks complete.
// You watch from the dashboard.
14
npm packages
122
tests passing
6
runtime adapters
12
agent personas
0
config required
< 1min
to first agent

Everything you need.
Nothing you don't.

14 packages that compose into any agentic product — CRM, support desk, devops, personal assistant.

Agent Execution

6 runtimes. 12 personas. 5 team templates. Org hierarchy with delegation and escalation. Create a full engineering team in one call.

Workflows

DAG engine with 6 block types. Wake agents, call connectors, branch on conditions. Smart routines that check before spawning.

Connectors

OAuth handled. Events typed. SDK for building your own. Slack + Google ship as reference. Community builds the rest.

Memory

Agents remember. Agents recall. Agents learn across runs. Pluggable — hebbs.ai, or bring your own cognitive engine.

Admin API

Full REST for everything. Agents, tasks, runs, approvals, budgets, routines, plugins. JWT + session auth. SSE realtime.

Headless UI

React hooks for everything. useAgents, useTasks, useRuns. No opinions on your markup. Build any dashboard.

From goal to done

You set the goal. The AI team figures out the rest.

1

Create team

One API call creates your AI workforce — CEO, CTO, engineers, QA — with hierarchy and personas pre-configured.

2

Assign a goal

"Build the MVP by Friday." Assign to the CEO or CTO. They understand what to do.

3

Delegation

CTO reads the goal, breaks it into subtasks, assigns each to the right engineer based on skills. QA gets the test plan.

4

Autonomous work

Each agent spawns a CLI (Claude, Codex, Gemini), writes code, runs tests, posts updates. Budget enforced. Memory shared.

5

Escalation & completion

Stuck? Agent escalates to its boss. Need approval? Humans approve via dashboard. Done? Tasks close, memory persists, next goal begins.

Zero config. Seriously.

One line boots the entire platform. Embedded Postgres included.

$ npx create-boringos my-startup
$ npm run dev
✓ Server running at http://localhost:3000
$ curl -X POST /api/admin/teams/from-template \
-d '{"template": "engineering"}'
✓ Created CTO (reports to: none)
✓ Created Senior Engineer (reports to: CTO)
✓ Created Engineer (reports to: CTO)
✓ Created QA Engineer (reports to: CTO)
$ curl -X POST /api/admin/agents/cto-id/wake \
-d '{"taskId": "build-mvp"}'
✓ CTO woken — reading task...
✓ CTO delegated "Auth module" → Senior Engineer
✓ CTO delegated "Dashboard" → Engineer
✓ CTO delegated "Test plan" → QA Engineer

Your AI team is working. Check the dashboard.

14 packages. Use what you need.

@boringos/coreApplication host
@boringos/agentExecution engine
@boringos/runtime6 CLI adapters
@boringos/memoryCognitive memory
@boringos/workflowDAG engine
@boringos/connectorConnector SDK
@boringos/driveFile storage
@boringos/dbPostgres + Drizzle
@boringos/pipelineJob queue
@boringos/uiReact hooks
@boringos/connector-slackSlack
@boringos/connector-googleGmail + Calendar
@boringos/sharedBase types
create-boringosCLI scaffold

The boring parts are handled.
Go build something exciting.

From zero to running agent platform in under a minute.

Read the docsView on GitHub