Nagent documentation
How to set up your workspace, hire your AI teams and get real work out of them, one screen at a time.
These are the platform docs, written for the people who use Nagent. Your developers can run agents from their own code, Claude Code or Cursor with the developer docs: the API, MCP, webhooks and the Python and TypeScript SDKs.
New to Nagent? Start here
- Step 1Signing upThe self-serve sign-up from the first screen to an open workspace, including choosing whole teams or single team leads, what each step asks for, the sign-in link, and what activates.Read
- Step 2Setting up with your team leadHow your team lead sets up your whole workspace in a conversation, saving each detail itself with Undo, leaving only tool sign-ins to you, with voice and video (Beta), check-ins, hand-overs, and how to switch to the set-up screens.Read
- Step 3Your first hourWhat Home shows a new workspace, how to read its counts, people, set-up ledger and agent list, and the order to do things in during the first hour.Read
Start here
- Getting startedWhat Nagent is, the few ideas you need before you begin, and which pages to read first to sign up, sign in and get through your first hour.
- How onboarding worksThe path from sign-up to your first approved piece of work in the first hour, then your first day, week and month, what Nagent does at each step, and where to read more.
- Signing upThe self-serve sign-up from the first screen to an open workspace, including choosing whole teams or single team leads, what each step asks for, the sign-in link, and what activates.
- Signing inThe three ways back into your workspace, the authenticator code some roles need, remembering a device and removing one, how a self-serve account gets a password, and what an invited teammate does the first time.
- On your phoneHow the workspace works on a phone: the team list as a drawer, the room's More menu and sheets, the Inbox one pane at a time, and approving on the move.
- Your first hourWhat Home shows a new workspace, how to read its counts, people, set-up ledger and agent list, and the order to do things in during the first hour.
- Setting up with your team leadHow your team lead sets up your whole workspace in a conversation, saving each detail itself with Undo, leaving only tool sign-ins to you, with voice and video (Beta), check-ins, hand-overs, and how to switch to the set-up screens.
- The Knowledge HubWhat the Knowledge Hub holds, its five sections and their tabs, the order to fill them in, what your agents read from it when they work, and the questions they could not answer.
- Connecting a toolHow to connect a tool such as Gmail, HubSpot or Slack, choose which of its actions your agents may call, narrow each action to the agents that need it, and keep the connection healthy.
- Staying in the loopThe emails Nagent sends about your workspace, approving from an email, the card that says what happened while you were away, making a team's job run every week, and how drafts learn from your edits.
- Sending feedbackHow to tell Nagent what broke, what was unclear or what you wish it did, the short questions on Home, and how you hear when it ships.
- Your reviewsThe review of your workspace at one month, one quarter and one year, what each figure counts, why it is set against your first week, and how to share a read-only link.
- GlossaryThe words Nagent's screens use, each in a sentence or two, with a link to the page that explains it properly.
- What's newWhat changed in Nagent at each production release, grouped by area and written in plain language, with a link to the page that explains each change.
Workspace
- Workspace set-upThe seven set-up cards, how each one is computed from what the workspace actually has, and the five questions the workspace has not answered yet.
- People and rolesHow to invite a teammate with a role at the whole organisation, a department or a team, what each of the five roles may do, how to grant and remove roles, and how to reset a password or deactivate someone.
- Approvals and the InboxWhat "queued for approval" means, where held work waits, how to approve or refuse it on the Actions page and in the Inbox, and who may decide what.
- Autonomy and governanceThe five autonomy levels, the six classes of action, how a level rises and drops, the governance profile, what always asks a person first, and the per-agent budget caps that stop an agent when it reaches them.
- Billing and plansHow to read the Billing and Plan page, what a new workspace with no plan runs on, how to buy, change or cancel a plan, how credits are metered, and how plans relate to the caps that stop agents.
- Credits and budgetsHow Nagent keeps agent spending inside limits you can see, what happens when an agent reaches its cap, and where to read what your workspace has spent.
- Agent voicesHow your team leads speak in voices of their own, how to have a team room read new messages aloud, how to talk to your lead in the set-up chat, how to dictate into any chat, and which voice engines a video persona can use.
Knowledge
- Who you sell toHow to describe your customer in your own words, have Nagent work out the targeting your agents use, and correct what it concluded.
- Products and servicesHow to record what you sell, why only active offerings reach your agents, and how to review the drafts Nagent extracts from your website, a page or a file.
- Bootstrapping from your websiteHow to point Nagent at your marketing site so your agents start with your own pages, which pages it reads first, what lands in the Knowledge Hub, and how to manage the sites you crawl.
- Brand Lock AIWhat your workspace's brand row is, how to draft and confirm your brand voice, add logos, colours and a guideline document, generate a brand book, and how agents and RUPA read it.
- Adding knowledgeHow to add a document to the Knowledge Hub by pasting text, from a URL or by uploading a file, how tags are suggested, and what the pending, embedded and failed statuses mean.
Teams
- Setting up a teamWhat a team holds on Nagent, where your teams are listed, and the two ways to add one, from a prebuilt team or from a blank charter, with the trust each agent keeps inside it.
- The team roomWhere people and agents work together. The room's tabs, the composer, threads, the rail beside the conversation, the agent surface, the Whole team room, and how rooms work on a phone.
- Workflows in a roomThe prebuilt workflows a team room carries, how to run one now or as a free test, keep it on its schedule, and follow each step as it runs.
- Long-running workHow a room shows each step while an agent works, how longer work runs as jobs you can follow, steer and stop, and the limits every job runs under.
- Building a workflowBuild a workflow from one sentence, review the drafted steps on the canvas, publish it, set what starts it, test it on a free model and cap what a run spends.
- Anatomy of a teamWhat a team is made of on Nagent: the conversation, the members list where people and agents stand side by side, and the files that hold the team's charter, tasks, decisions and memory.
- The team filesThe files behind every team room, from the charter and roster to memory, the board, decisions, shared documents, run files and the events log, and how people and agents read them.
- Extending a teamHow to add people and agents to a team from its member list, what changes when someone joins or leaves, and how to grow a team without flooding the approval queue.
- Sending a briefHow work starts in a team room, the difference between asking a question and giving an instruction, how an instruction becomes a task on the board, and where the result comes back.
- Meetings in the team roomFollow one meeting in a team room, from the lead calling it or you starting a huddle, to the tasks it leaves on the board, and what you can do at each step.
- HuddlesStart a meeting in a team room yourself with the Huddle button. The lead runs it as a working session, writes the tasks, and can add to the team on its own word.
- Building a team of your ownA worked example. Build four agents, give them Reddit and skills, form a team, add a weekly workflow, and ask the team for a plan in its room.
Agents
- What is an agentWhat an agent is on Nagent, the parts every agent is made of, when a custom agent is worth building, and what Nagent adds that an agent builder does not.
- Creating an agentHow to build a custom agent from one sentence or a template, answer the builder's questions, review the draft, provision it live, or build a team from the chat.
- Agent WorkbenchHow to read the Agent Workbench, its level ladder and org chart, and each agent's own page, where its instructions, tools, triggers, guardrails, budget and status are set.
- Boards and surface viewsWhat an agent writes to the AI Surface, how the team's board shows it as a list, a calendar, a checklist or KPI cards, and which types each agent may write.
- SkillsWhat a skill is, how to read the Skills page, how a skill is written, who may create one, how to attach one to an agent, and how skills sit beside levels and tool access.
- MIRA: the coordinatorThe places MIRA appears across Nagent, what she leads in your workspace, what the Ask MIRA panel on Home can and cannot answer today, and how her spend is counted.
Tools
- Third-party toolsThe kinds of outside tools a workspace can connect, how an agent uses a connected tool during its work, and the checks every call passes before it runs.
- Email and integrationsHow to send your workspace's outbound email from your own domain with a Resend key, what happens until you do, and what else lives on the Integrations settings page.
- An agent's own email addressHow to give an agent its own address on agents.nagent.ai, what happens when someone writes to it, who gets an answer with tools and who without, how its replies are approved and sent, and the rules that keep automatic mail away from strangers.
- Triggers and webhooksHow outside services start work in your workspace through signed webhooks, how MCP tokens let an approved client call Nagent, and how event subscriptions decide what fires an agent.
- Nagent for SlackHow to install the Nagent app in Slack, what it reads from your public channels, ask your team or an agent in a channel or a direct message, link your Slack account, approve an agent's reply in Slack, and disconnect.
Prebuilt teams
- Prebuilt teamsThe three teams Nagent prebuilds, who each one seats and who leads it, how the sign-up cards differ from what installs, how to add a team after sign-up, and the Teams to add offered beside them.
- The Sales teamWho arrives when you choose Sales at sign-up, what each of the five AI teammates owns, how work passes between them, and what the team cannot do yet.
- NORA: outbound research and activationHow NORA finds and researches accounts, drafts first-touch outreach, hands a qualified meeting to a person, and which of her actions wait for approval.
- Running outbound campaigns with NORAHow to save an audience as a campaign on NORA's Discover screen, set how and when it runs, read what each run found, and take the companies it finds through to an approved first e-mail.
- DEXA: deal executionWhat DEXA owns once a human takes a deal, what it reads and writes, which of its actions ask first, and how to brief it from the room or its own screen.
- Order forms and contractsHow DEXA drafts an order form from a lead, how a person reviews, approves and sends it, and how the buyer reads and accepts it in the deal room without signing in.
- RIVA: revenue operationsHow RIVA keeps the pipeline true, what it measures and flags, why it can never edit a deal, and how to ask it questions.
- LIPI: CRM email writerWhat LIPI writes, what it grounds every draft in, where its drafts appear, and who decides whether they are sent.
- VANI: outbound email writerWhat VANI writes for NORA, what it grounds a cold email in, and where the decision to send actually sits.
- Unsubscribe and suppressionHow a prospect opts out of your outbound email, what the suppression list records, and how every send path and the sequencer honour it from then on.
- NIA: performance marketingWhat NIA reads in your Google Ads account, what she proposes, why every live change to your ads waits for your approval, and where to work with her.
- NIVA: investor relationsWhat NIVA keeps for a fundraise, which calls stay yours, how the Investor Relations room works, and what an investor sees in the deck and the data room.
Programmes
- DRIS, AI VisibilityHow DRIS measures your company in search and in the answer engines, ranks fixes against the objective you declare, and turns a proposal you approve into a change on your site.
- Setting up DRISConnect Search Console, Google Analytics and crawler logs, name your competitors and prompts, and check each DRIS screen is reading real data before you rely on it.
- HOOK, the ad creative team leadHow HOOK turns a paid social brief into angles, hooks and scripts, checks every claim before anything ships, and learns from how your ads performed.
- MOXA, socialHow MOXA plans a week of social posts from your programme, drafts and checks each one, holds it for a second person to approve, listens for what comes back, and measures what worked.
- CREA and the content studioHow CREA's team writes blog posts, decks, documents and landing pages grounded in your Knowledge Hub and brand, how you review and publish them, and how to ask for them from any room.
- Decks and landing pagesHow PAIGE drafts decks and PDF documents you edit and download as PDF or PowerPoint, and how LANDON builds landing pages you publish, with leads landing in your CRM and the page served on your own domain.
- RUPA, the image studioHow RUPA renders images to a preset with your brand attached, when a render waits for your approval, what it costs, and where the finished plates are kept.
- Echo, the reputation coworkerHow Echo reads your Google Business reviews into cases, samples what answer engines say about you and checks each claim against your approved facts, and what still waits on a person.
- ALPHA, the video studioHow to describe a video to ALPHA, pick the desk that makes that kind of video, read the plan and quote in credits before anything renders, approve the storyboard, follow the job through its eight stages, and read Personas, Mission control and Routes.
- ALPHA personas and avatarsHow to make the spokespersons and digital twins ALPHA puts on screen, from a still, from words or from footage, with their voices, releases, consent and approval.
- VIDYA, long-form video for your learning seriesHow VIDYA turns an approved script into a finished long-form episode, presented by you, an AI presenter or both, and how you download it, post it, cut it into shorts or embed it.
- CORA's CRMYour leads and pipeline in Nagent, how agents read and write them, how email to a lead is written and sent from a rep's own mailbox, and where agent work waits for your approval.
- MeetingsThe two kinds of meeting in Nagent, meetings with your customers on the lead record and meetings your agents hold with each other, and where each one is set up.
Developers
- Connect your siteHow your website reads the search and AI visibility fixes your team approves in Nagent, your published blog posts and your landing pages, with one server-side key.
- The API and MCPHow to create an API key, run your agents from your own code or from a coding agent such as Claude Code or Cursor, follow a run, receive signed webhooks, and keep keys safe, with the same autonomy levels and approvals as inside Nagent.