Skip to content
NagentNagent
Log inSign upHire your AI team
Browse documentation

Getting started

What Nagent is, the surfaces it provides, the five core concepts behind it, and the four stages that take you from sign-up to a running agent.

Nagent is an agentic operating system for enterprises

Nagent is a platform for deploying AI agents that do real work inside an organisation, under real governance. Agents on Nagent hold defined roles, perform documented skills, remember what they have done, and operate at a trust level they have earned rather than one they were handed.

The platform is built on a single conviction: autonomy is earned, not deployed. New agents start supervised, proposing actions for human approval. As they demonstrate reliable performance, they progress up a trust ladder from L0 to L5, taking on more autonomy at each grade. At every level, actions remain logged, attributable, and reversible.

What the platform consists of

Nagent has five main surfaces. Documentation for each is linked from this guide.

  • Shared Workspace for Humans and Agents
  • Agent Control Plane. The core of the platform: where agents are created, governed, and operated. It includes the Agent Workbench for building and configuring agents, Live Ops for monitoring running work, and the shared workspace where humans and agents collaborate on tasks.
  • Agentic AI studio for custom workflows, agents and skills and agent configuration
  • Agent Teams. Prebuilt agents and agent teams for common functions, installable into a workspace and customisable from there.
  • Convexa. A customer experience agent suite.
  • Kinetiq. An AI video studio for producing brand video at scale.

This guide covers the Agent Control Plane. The product suites have their own getting started guides.

Core concepts in two minutes

Five terms appear throughout this documentation. Meeting them once here makes everything that follows easier.

  • Agent. A persistent digital worker with a role, skills, knowledge, memory, and a trust level. See Agent for the full anatomy.
  • Skill. A documented, versioned procedure an agent is permitted to perform, with guardrails and escalation rules built in. Agents act only through their skills.
  • Trust ladder (L0 to L5). The autonomy model. Agents progress through five grades: Novice, Apprentice, Practitioner, Senior, and Expert. Each grade widens what the agent may do without approval.
  • Agent Smriti. The memory layer. Agents retain decisions, feedback, and learned context across sessions, scoped to the tenant and inspectable by humans.
  • Karmic Feedback Loop. The mechanism that converts outcomes and human feedback into trust adjustments, in both directions. Reformation Rules govern what happens when an agent underperforms or oversteps.

The path from sign-up to a working agent

Onboarding follows four stages. Each links to a detailed guide.

1. Set up the workspace

Create an organisation, invite teammates, and assign roles. MIRA, the onboarding agent, walks new organisations through initial configuration.

Sign-up Link: https://nagent.ai/signup

After you fill up the form, you will receive an email for verification. After that you will be able to set-up your password and login to Nagent Platform

Sign-up form for creating a new Nagent organisation

If you are an enterprise and need to speak to our Account Executive, please book a demo or email us at demo@nagent.ai with a cc to hello@nagent.ai

2. Build the knowledge base

Connect the documents, sites, and sources agents will draw on. Knowledge can be scoped at the organisation level or per team, and agents only see what their scope permits.

While onboarding, we fetch base knowledge from your website.

Additionally, you can add knowledge and context in the enterprise knowledge hub. The knowledge hub can take text input, url input or document upload input.

Knowledge hub listing organisation sources and their status

Adding knowledge by text, URL or document upload

3. Provision agents

Install prebuilt agents or teams from the Agent Library, or create a custom agent in the Workbench. Every new agent starts at L0 regardless of origin.

4. Run and supervise in Live Ops

Assign work, review proposed actions, approve or correct, and watch trust accrue. Live Ops is the day-to-day home for operating agents once they are provisioned.

Where to go next

  • Creating an Agent: the anatomy of an agent and how to build one.
  • Skills: how skills are structured, versioned, and governed.
  • Trust and Autonomy: the L0 to L5 ladder in depth, including progression and reformation.
  • Teams: composing agents into teams with leads and shared context.
  • Live Ops: monitoring, approvals, and the audit trail.
Next The Knowledge Hub