# DutyGraph Docs > DutyGraph is the AI-readable operating model of a company: who does what, in Task Cards, role profiles and a Company Repo. These docs explain the product, how advisors and forward-deployed engineers run an engagement, and how to connect over MCP and the CLI. Every page is also served as plain Markdown: add .md to its address. llms-full.txt holds the whole site in one file. Pages label each feature Built, In review or Planned; never describe a Planned item as available. ## Table of Contents ### Start - [How it all fits together](/start/how-it-fits-together.md): One page that shows how every part of DutyGraph connects: the people, the screens, the records and the AI surfaces. - [Words we use](/start/words-we-use.md): The short list of DutyGraph terms, with the meaning each one has in the product. - [Built and planned](/start/built-and-planned.md): What is built, what is in review and what is planned, for the features these docs describe. ### Advisor dashboard: the manual - [Run an engagement](/operate.md): The operating manual for advisors and forward-deployed engineers: the six steps of a DutyGraph engagement and what you do in each. - [1. Pitch to Package](/operate/pitch-to-package.md): Step 1 of an engagement. Research the company, plan the call, run it, and send the onboarding package. - [2. North Star session](/operate/north-star-session.md): Step 2 of an engagement. Load the team, run the top-level session with the sponsor, and save the company's vision, business model and goals. - [3. Org chart and role profiles](/operate/org-chart-and-role-profiles.md): Step 3 of an engagement. How role profiles are drafted from the top down, what is in one, and how they are approved. - [4. The client app](/operate/client-app.md): Step 4 of an engagement. What the client's team sees in their app, how they sign in, and how to help someone who is stuck. - [5. Task Cards](/operate/task-cards.md): Step 5 of an engagement. What a Task Card holds, how people write them, and how you check them. - [Work Map and workflows](/operate/work-map.md): The Work Map in the advisor dashboard: the Process map, Work flows with suggested workflows, what to fix next, asking a person, and DutyGraph Agent finding the constraint. - [Data formats and the hand-off format fixer](/operate/data-formats.md): Record the format of what comes into and goes out of each Task Card, find where two formats do not match, save a reusable conversion, build an agent from it and run it from the app or an AI chat. - [6. Readout](/operate/readout.md): Step 6 of an engagement. The Readout is the report that shows the finished map and what it makes possible. - [Role AI](/operate/role-ai.md): The AI persona built from each approved role profile, with first questions and task starters, and how it is used for coaching, onboarding and training. - [Company Repo and DutyGraph Agentâ„¢](/operate/company-repo-and-dutygraph-agent.md): How the Company Repo is laid out, what you can edit, and how DutyGraph Agent writes documents for you. - [Keep it current](/operate/keep-it-current.md): How a mapped company stays current after the Readout, with Autopilot passes that find gaps and propose questions. - [Partner firms](/operate/partner-firms.md): How a coaching firm or an AI firm runs DutyGraph for its own clients, with its own workspace, MCP and the CLI. - [Your workspace: sign-in, password and invitations](/operate/workspace-and-invitations.md) - [Trust and privacy](/operate/trust-and-privacy.md): What employees' answers are used for, who can see what, and the settings that control it. ### Client app: guide for the team - [The client app](/client-app.md): A guide for a client's team: what the DutyGraph client app is, how to sign in, and the main screens on a laptop and a phone. - [Answer your questions](/client-app/answer-questions.md): How a team member answers their questions in the DutyGraph client app: Home, the Do first cards, a question set, typing or speaking an answer, and saving a section. - [My work, Files and your company](/client-app/my-work-and-files.md): My work (Talk about your work, DutyGraph Agent, Mentioned you, decisions from your advisor), Your place, Files, your company's direction, Strategy, Improve how work flows and Team in the DutyGraph client app. - [Connect your AI](/client-app/connect-ai.md): The Connect AI screen in the DutyGraph client app: connect the AI you already use, or copy a prompt into it and paste the answer back as draft Task Cards. - [On your phone](/client-app/on-your-phone.md): The DutyGraph client app on a phone: Home, answering a question, the More menu, Add, My work, Files and Connect AI at phone size. ### Playbooks - [Playbooks](/playbooks.md): Short recipes for the jobs an advisor or FDE does most often. - [Analysis pages from the map](/playbooks/analysis-pages.md): An idea for a playbook where your AI turns a company's map into small web pages with charts, frameworks and forecasts. ### Offers - [Knowledge transfer for M\&A](/offers/knowledge-transfer-for-m-and-a.md): An offer for buyers and sellers of companies. Pay the people who hold the know-how to write it down fast, so the knowledge transfers in weeks instead of a year. ### Build: MCP and CLI - [Quickstart](/build/quickstart.md): Connect Claude Code to your DutyGraph workspace over MCP and read your first company in five minutes. - [Sign-in and tokens](/build/authentication.md): The two ways to sign in to DutyGraph's MCP server and CLI, the scopes a token can have, and how to limit a token. - [Connect an AI assistant](/build/mcp.md): Connect Claude, ChatGPT, Cursor or any MCP client to DutyGraph's MCP server. - [Employee tools](/build/employee-tools.md): What an employee's own AI chat can do when it is connected to DutyGraph, and the interview rules it follows. - [CLI](/build/cli.md): Install the dutygraph CLI, sign in, and run your first commands. - [Errors and limits](/build/errors-and-limits.md): The errors and limits you meet most often over MCP and the CLI, and what to do about each. - [REST API](/build/rest-api.md): Why there is no public REST API with keys today, and what to build on. ### Reference - [MCP tools](/reference/mcp-tools.md): Every DutyGraph MCP tool, its scope and what it does, read from the server code. - [CLI commands](/reference/cli-commands.md): Every dutygraph CLI command with examples and the rules it follows. - [MCP and CLI for partners](/reference/partners.md): What a partner firm can do with a client through DutyGraph's MCP server and CLI today, with every endpoint, tool and command. ### Training - [Training path](/training.md): The training path for a new advisor or forward-deployed engineer, in eight modules with a check at the end of each. - [AI trainer](/training/ai-trainer.md): How to use an AI assistant as a trainer for new advisors and FDEs, with these docs as its source. - [Test flows](/training/test-flows.md): Seven test flows to check a new DutyGraph workspace end to end: advisor dashboard, client app, CLI and MCP, with the expected result at each step. ### For AI - [AI-readable docs](/ai.md): How AI assistants and agents read these docs: llms.txt, llms-full.txt and a Markdown copy of every page. ### Product handbook - [DutyGraph](/handbook.md): DutyGraph maps how a company actually works, finds where AI can help, and governs the authority of the agents built on that work. #### Start here - [Start here](/handbook/start-here.md): What the Start here section covers, and which pages to read first for your role. - [What is DutyGraph](/handbook/start-here/what-is-dutygraph.md): DutyGraph in plain English: an evidence-backed model of how a company works, and the delegated authority of the AI agents built on that work. - [Three ways in, one model](/handbook/start-here/three-ways-in.md): DutyGraph can start from company strategy, from one department, or from one agent request. All three fill the same operating model. - [The DutyGraph loop](/handbook/start-here/the-dutygraph-loop.md): The six-stage loop DutyGraph is designed around, which parts are built today, and where each one lives in the app. - [Run a company in 9 steps](/handbook/start-here/run-a-company.md): How an advisor takes one company from its website to a governed agent, in 9 steps, with the status of each step. - [Who uses DutyGraph](/handbook/start-here/who-uses-dutygraph.md): Each role that touches DutyGraph, what they see and do, and how they get in. - [Built and planned](/handbook/start-here/built-and-planned.md): The maintained public status reference for DutyGraph. What is in the code today, what is simulated and what is planned. #### Map the company - [Map the company](/handbook/discovery.md): How DutyGraph maps a company, the three ways capture can start, and what each step adds to the Company Repo. - [Company research](/handbook/discovery/company-research.md): Add a company by its website, research it, and keep the research current without losing what people confirmed. - [Sponsor packet](/handbook/discovery/sponsor-packet.md): The first ask to the client contact, what they see on their private page, and what comes back to the advisor. - [Executive kickoff](/handbook/discovery/executive-kickoff.md): Plan, schedule and record the leadership meeting, and what it adds to the Company Repo. - [Team asks](/handbook/discovery/team-asks.md): Send short asks to each person, how they answer on a private link, and how the advisor reviews what comes back. - [Company Repo](/handbook/discovery/company-repo.md): The company as files an AI can read, with a state on every file. - [Task Cards](/handbook/discovery/task-cards.md): One piece of recurring work: what a Task Card holds, how it becomes evidence, and how it relates to an agent's permissions today and in the target design. - [Talk first, then three parts](/handbook/discovery/talk-first-capture.md): How a person describes their work by talking, how the AI sorts what they said into tasks with three parts, and the reasoning steps the AI follows. The same method in an AI chat and in the client app. - [Work Map](/handbook/discovery/work-map.md): How the Work Map shows the flow of work, who owns and does each Task Card, and where the gaps are. #### Find opportunities - [Find opportunities](/handbook/opportunities.md): How DutyGraph turns a mapped company into improvement and AI opportunities, and what is Built today. - [Opportunities](/handbook/opportunities/opportunities.md): How DutyGraph shows each person's hours, proposes AI modes and agents by code, and lets a manager suggest an agent to an employee. - [Outcomes and Strategy](/handbook/opportunities/outcomes-and-strategy.md): How DutyGraph finds possible constraints, runs strategy frameworks on the Company Repo, and tests a change with measures and pilots. - [Route](/handbook/opportunities/route.md): How Route finds the cheapest change that can work for each accepted possible constraint, sizes it in hours and plans its proof. - [Reports](/handbook/opportunities/reports.md): How the Reports tab builds dated reports from the Company Repo, and how an advisor saves, approves and delivers them. #### Build agents - [Build agents](/handbook/agents.md): How an opportunity becomes an agent manifest, and what DutyGraph does and does not do with it. - [Agent builder and manifests](/handbook/agents/agent-builder.md): The agent builder workspace, how each field is filled and limited, and what a saved manifest holds. - [Share the agent plan](/handbook/agents/agent-plan.md): Send the person an agent would act for a private, plain plan, and record their answer. - [Exports and runtimes](/handbook/agents/exports-and-runtimes.md): Every file DutyGraph writes for an agent, which runtime each is formatted for, and when DutyGraph refuses. #### Govern agents - [Govern agents](/handbook/governance.md): How DutyGraph proposes what an agent may hold, who approves it, how each decision is recorded, and where your own identity systems take over. - [The key join](/handbook/governance/the-key-join.md): How DutyGraph joins Task Cards, the owner's holdings, the System Catalog and the rule packs into an agent's proposed scopes, what the evaluator checks today, and the gap that remains. - [Why governed agents](/handbook/governance/why-governed-agents.md): Agents are getting access faster than companies can explain why they need it. - [Holdings](/handbook/governance/holdings.md): The access a person already has, one scope at a time, with its source and date. - [System Catalog](/handbook/governance/system-catalog.md): The scopes each business system offers, in one grammar, and why the default scopes are coarse. - [Rule packs and SoD](/handbook/governance/rule-packs-and-sod.md): Agent rules, segregation-of-duties rules and mitigating controls, stored as versioned packs. - [Approval chain](/handbook/governance/approval-chain.md): The people who approve human access decide on agent access. Three decisions, every step. - [Ledger](/handbook/governance/ledger.md): A hash-linked, append-only record of each governance decision, and what its integrity check does and does not show. - [Agent gateway](/handbook/governance/agent-gateway.md): What the DutyGraph gateway checks on each call an issued agent routes through it, what it records, and what it does not cover. - [Approval, provisioning and revalidation](/handbook/governance/approval-provisioning-and-revalidation.md): What DutyGraph decides and records, what your identity systems provision, what runtimes enforce, and how an approval should be checked again when the company changes. - [Set up governance](/handbook/governance/set-up-governance.md): Load the six packs that turn on governed agents for a company. - [Request a governed agent](/handbook/governance/request-a-governed-agent.md): From a request to an issued agent, with each decision recorded, and what issue does and does not do. - [Request from your AI](/handbook/governance/request-from-your-ai.md): Ask for an agent from the AI chat you already use. Approval happens in DutyGraph. - [Audit evidence](/handbook/governance/audit-evidence.md): What population DutyGraph can show an auditor, how a sample is traced from grant to gateway call, and what reconciliation with the wider enterprise still needs. #### Run engagements - [Run engagements](/handbook/advisors.md): How an advisor's workspace is laid out, and how an engagement runs from first note to Living DutyGraph. - [Accounts and Inbox](/handbook/advisors/accounts-and-inbox.md): How an advisor tracks client companies, reaches out within DutyGraph's outreach rules, and works the Inbox. - [Engagements and payment](/handbook/advisors/engagements-and-payment.md): How an advisor sends a client their options, how payment or a recorded commitment activates the engagement, and how a company is shared. - [Living DutyGraph](/handbook/advisors/living-dutygraph.md): What happens after the audit, with digests, research refreshes, renewals and client results that each wait for the advisor. - [DutyGraph Agent](/handbook/advisors/repo-copilot.md): The AI panel beside the Company Repo that answers, reviews, proposes edits and drafts asks, and only ever proposes. - [Phones and notifications](/handbook/advisors/phones-and-notifications.md): What changes at phone width, how advisors get phone notifications, how to share private links, and where the Companion phone app stands. #### Connect - [Connect](/handbook/connect.md): The ways AI tools and outside services connect to DutyGraph, and which are built or planned. - [MCP for advisors](/handbook/connect/mcp.md): Connect an advisor's AI assistant to a DutyGraph workspace over MCP, with scoped personal tokens. - [Integrations](/handbook/connect/integrations.md): The outside services DutyGraph connects to today, how each connects, and which connections are planned. #### Architecture - [Architecture](/handbook/architecture.md): How DutyGraph fits together, for a technical buyer or evaluator, and where to read more. - [System architecture](/handbook/architecture/system-architecture.md): DutyGraph's components and trust boundaries, the three entry points into one model, the path from discovery to authority, the Company Repo, tenancy, the MCP surfaces and hosting. - [Data model](/handbook/architecture/data-model.md): The main things DutyGraph records about a company, how they relate, and how versions, content hashes and stable ids keep them traceable. - [How DutyGraph uses AI](/handbook/architecture/how-dutygraph-uses-ai.md): Every place DutyGraph calls an AI model, every place code decides instead, which key each call uses, and the rules all of them follow. - [The role AI persona](/handbook/architecture/role-ai-persona.md): How DutyGraph builds the AI persona for a role from its approved role profile, what each block of the prompt does, and why this helps an AI do the role's work. #### Trust and security - [Trust center](/handbook/trust.md): DutyGraph's trust center. How DutyGraph protects each company's data, where it stands on SOC 2, who processes data for it, how long it keeps data, and how to reach us. - [Security and data handling](/handbook/trust/security-and-data.md): What the code does to keep companies apart, check permissions, store secrets, limit what reaches AI providers and retain data, and what is not in place yet. - [SOC 2 status](/handbook/trust/soc-2-status.md): Where DutyGraph stands on SOC 2: no report yet, readiness work started, and what is not in place. - [Subprocessors](/handbook/trust/subprocessors.md): The outside services that process data for DutyGraph: what each one does, what it receives, and when it is used. - [Data retention](/handbook/trust/data-retention.md): How long DutyGraph keeps each kind of data by default, what Erase person removes, and what DutyGraph does not delete yet. - [Evidence and provenance](/handbook/trust/evidence-and-provenance.md): How DutyGraph keeps research, AI proposals, a person's confirmation, the advisor's approval and outside authority apart, and how it records where each fact came from. #### Roadmap - [Roadmap](/handbook/roadmap.md): Planned work, in order but without dates, and the long-term direction. - [Just-in-time task capture](/handbook/roadmap/just-in-time-task-capture.md): Planned: when an employee requests an agent and no adequate Task Card exists, capture just that one task, then govern the request as usual. - [Employee Knowledge Compensation](/handbook/roadmap/employee-knowledge-compensation.md): How DutyGraph proposes to measure and pay employees for the work knowledge they put into Task Cards, and how it protects them and the company. - [Workflow Gainsharing](/handbook/roadmap/workflow-gainsharing.md): How DutyGraph proposes to measure the time and output an employee gains on the Task Cards they run, and how a company can share that gain with the employee. - [The workflow coach and the AI use check](/handbook/roadmap/workflow-coach.md): A proposed AI coach in the employee's app that asks which AI a person uses for each task, and proposes a faster, cheaper or safer way to do the work. - [Always-on DutyGraph Agent](/handbook/roadmap/always-on-agent.md): How DutyGraph proposes to let the DutyGraph Agent keep working between visits: wake on a schedule or an event, check the map, and send a few approved questions. - [Key-person risk map](/handbook/roadmap/key-person-risk-map.md): An optional add-on an advisor can offer after mapping: a page, built from the company's Task Cards, that shows who the business depends on, with a link the owner can forward to a buyer or banker. - [Private deployment and private AI](/handbook/roadmap/private-deployment.md): An optional add-on service for clients who want DutyGraph in their own cloud account, on their own AI contract, on a model they host, or fully offline. Four tiers, what is built and what each needs. - [Where DutyGraph is going](/handbook/roadmap/vision.md): The long-term direction for DutyGraph, and which parts of it exist today. #### Reference - [Reference](/handbook/reference.md): The terms these docs use and answers to common questions. - [Glossary](/handbook/reference/glossary.md): The terms these docs and the app use, each with one meaning. - [FAQ](/handbook/reference/faq.md): Short, honest answers to the questions buyers, IAM teams, auditors, employees and advisors ask. ### Advisor guide (in-app Help) - [Advisor guide](/advisor-guide.md) - [Hosted pilot: start here](/advisor-guide/00-hosted-quickstart.md) - [Start here: what Duty Graph does](/advisor-guide/01-start-here.md) - [Guided walkthrough: your first engagement](/advisor-guide/02-walkthrough.md) - [Full user manual](/advisor-guide/03-user-manual.md) - [Advisor playbook: execute an engagement A to Z](/advisor-guide/04-advisor-playbook.md) - [Training workbook](/advisor-guide/05-training-workbook.md) - [Facilitator guide and answer key](/advisor-guide/06-facilitator-guide.md) - [Client deliverables and example outputs](/advisor-guide/07-client-deliverables.md) - [Glossary and quick reference](/advisor-guide/08-glossary.md) - [Explore the Cobalt guided example](/advisor-guide/19-cobalt-guided-examples.md) - [Read a task card in seconds](/advisor-guide/20-task-library.md) - [From evidence to a better result](/advisor-guide/21-strategy-and-weekly-review.md) - [Strategy: possible constraints, frameworks and DutyGraph Agent](/advisor-guide/22-strategy-reports-and-copilot.md) - [From Company Research to confirmed work](/advisor-guide/23-discovery-to-confirmed-work.md) - [Live framework canvases](/advisor-guide/24-live-framework-canvases.md) - [Answer a work request from your private link](/advisor-guide/25-participant-review.md) - [Advisor starter program: practice, review, then lead](/advisor-guide/26-advisor-starter-program.md) - [Classify the business and shape its operating flows](/advisor-guide/27-business-types.md) - [Work Map](/advisor-guide/28-company-work-map.md) - [The Company Repo and DutyGraph Agent](/advisor-guide/29-company-repo-and-copilot.md) - [Agents: proposed agents and agent manifests](/advisor-guide/30-agents-and-manifests.md) - [Route: the change for each possible constraint](/advisor-guide/31-route.md) - [Opportunities: your team's hours and the proposed agents](/advisor-guide/32-opportunities.md) - [Run a company through DutyGraph](/advisor-guide/33-run-a-company.md) - [Client app: Talk, DutyGraph Agent and one-tap checks](/advisor-guide/34-client-app-talk-and-agent.md)