> ## Documentation Index
> Fetch the complete documentation index at: https://docs.useorgx.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# OrgX

> OrgX is proof for AI-delivered work: a shared operating record for initiatives, decisions, artifacts, approvals, and receipts across the AI tools your team already uses.

## What is OrgX?

OrgX is **proof for AI-delivered work** — the operating record your agents and tools share. It connects the AI clients you already use to the same initiatives, decisions, artifacts, approval boundaries, and receipts so work can continue across sessions without losing context or human control.

**Supported clients**: Claude, Cursor, Codex, VS Code, Windsurf, Zed, ChatGPT, Gemini CLI

## Questions This Page Answers

* Is OrgX just another generic assistant, or an execution system?
* Which agent domains actually exist today?
* Why would a founder or operator pick this over ad-hoc prompting?
* What does the human still control?

<CardGroup cols={2}>
  <Card title="For Solo Founders" icon="user" href="/docs/quickstart">
    Keep decisions, artifacts, and next actions intact when work moves between
    chats, tickets, repos, and AI clients.
  </Card>

  <Card title="For Small Teams" icon="users" href="/docs/quickstart">
    Give teammates and connected agents one reviewable record of what is in
    progress, what was approved, and what evidence supports completion.
  </Card>
</CardGroup>

## The Problem We Solve

AI work fragments quickly: one client has the plan, another has the code, a
third has the approved decision, and the final artifact arrives without a
durable record of how it was produced. OrgX gives that work a shared spine.

| Without OrgX                        | With OrgX                                        |
| ----------------------------------- | ------------------------------------------------ |
| Context is rebuilt in every session | Connected clients orient from shared state       |
| Decisions disappear into chat logs  | Decisions remain attached to the work            |
| “Done” is a status update           | Completion can carry a receipt and evidence      |
| Autonomy is implicit                | Tool access and approval boundaries are explicit |

## How It Works

<Steps>
  <Step title="Connect Your Stack">
    Start in OrgX first, then add GitHub, Linear, Slack, and other tools via
    OAuth or MCP connectors only when you want them in the loop.
  </Step>

  <Step title="Agents Get to Work">
    Specialized agents (Engineering, Product, Marketing, Sales, Design,
    Operations, Orchestrator) analyze work items and execute against them.
    Agents can draft specs, open PRs, coordinate launches, and keep execution
    moving.
  </Step>

  <Step title="You Stay in Control">
    Work that reaches a configured judgment boundary becomes a **decision** in
    Mission Control. Approve, reject, or request changes with the relevant
    context attached.
  </Step>

  <Step title="Artifacts Ship">
    Work becomes artifacts—docs, PRs, campaigns, and reports—with the decisions,
    approvals, citations, and verifier results available for that run.
  </Step>
</Steps>

## Access OrgX Everywhere

<CardGroup cols={3}>
  <Card title="Web App" icon="browser" href="https://useorgx.com">
    Full Mission Control with canvas, decisions, and agent console.
  </Card>

  <Card title="ChatGPT" icon="comments" href="/docs/guides/chatgpt-integration">
    Approve decisions and manage agents directly in ChatGPT.
  </Card>

  <Card title="AI Toolkit" icon="toolbox" href="/docs/guides/toolkit">
    Connect Claude, Cursor, Codex, VS Code, ChatGPT, and more via wizard or MCP.
  </Card>
</CardGroup>

## Core Concepts

<AccordionGroup>
  <Accordion title="Agents" icon="robot">
    Specialized AI workers that handle specific domains:

    * **Engineering Agent**: Code changes, PR creation, bug fixes
    * **Product Agent**: PRDs, strategy, prioritization, research synthesis
    * **Marketing Agent**: Campaign briefs, content creation, SEO
    * **Sales Agent**: Lead qualification, outreach sequences
    * **Design Agent**: UX rationale, design brief assistance
    * **Operations Agent**: Incident/process operations and runbooks
    * **Orchestrator Agent**: Cross-domain routing and dependency management
  </Accordion>

  {' '}

  <Accordion title="Artifacts" icon="file-lines">
    Outputs from agent work—specs, PRs, campaign briefs, and research reports—kept
    with their available sources, verifier results, approval state, and execution
    history.
  </Accordion>

  {' '}

  <Accordion title="Decisions" icon="check-circle">
    Points where agents need human judgment. Every decision includes: - Context
    and reasoning from the agent - One-click approve/reject with optional notes -
    Automatic execution once approved
  </Accordion>

  <Accordion title="Mission Control" icon="compass">
    Your operational dashboard showing:

    * Active initiatives and their health
    * Pending decisions requiring your attention
    * Agent activity and work in progress
    * Artifacts ready for review
  </Accordion>
</AccordionGroup>

## What Makes OrgX Different

<CardGroup cols={2}>
  <Card title="Evidence-Based" icon="magnifying-glass">
    Artifacts can preserve citations, verifier results, approval state, and the
    execution receipt behind the work. Evidence stays inspectable instead of
    being reduced to a confidence claim.
  </Card>

  <Card title="Human-in-the-Loop" icon="hand">
    Agents propose, you approve. Configurable autonomy levels let you dial in
    exactly how much oversight you want.
  </Card>

  <Card title="Optional Integrations" icon="plug">
    Add GitHub, Linear, Slack, Notion, and more when your workflow needs them.
    Agents work with your existing stack, not around it.
  </Card>

  <Card title="MCP-First" icon="server">
    Built on the Model Context Protocol. Connect OrgX to any MCP-compatible
    client—ChatGPT, Claude, Cursor, and beyond.
  </Card>
</CardGroup>

## Why Teams Choose OrgX

* **Shared record, not another chat wrapper**: Connect supported AI clients through MCP or the wizard CLI.
* **Execution over chat**: Initiatives, decisions, artifacts, and receipts remain linked instead of being lost in threads.
* **Proof over promises**: Decisions and artifacts carry evidence, receipts, and attribution.
* **Control without micromanagement**: You approve high-leverage checkpoints, not every micro-step.
* **Domain specialists, not one blob model**: Routing picks the right domain agent for the job.

## Ready to Start?

<CardGroup cols={2}>
  <Card title="Quickstart" icon="bolt" href="/docs/quickstart">
    Get OrgX running in under 10 minutes.
  </Card>

  <Card title="ChatGPT Integration" icon="comments" href="/docs/guides/chatgpt-integration">
    Connect OrgX to ChatGPT for conversational control.
  </Card>
</CardGroup>