> ## 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.

# FAQ

> Common questions about OrgX: what it is, who it is for, how it differs from using AI clients alone, trust and approval workflows, pricing, and current capabilities.

## General

<AccordionGroup>
  <Accordion title="What is OrgX?">
    OrgX is an AI-powered operational platform that deploys specialized agents
    to handle coordination, decisions, and execution across your organization.
    It can integrate with tools like GitHub and Linear today, with Slack and
    other connectors planned. Integrations are optional. OrgX still keeps you in control through a
    human-in-the-loop approval system.
  </Accordion>

  <Accordion title="Who is OrgX for?">
    OrgX is designed for:

    * **Solo technical founders** who spend too much time on coordination
    * **Small teams (1-10 people)** who need operational leverage without hiring
    * **Product-minded leaders** who want AI assistance with planning and execution

    Our sweet spot is pre-seed to seed-stage companies doing \$0-50K MRR.
  </Accordion>

  {' '}

  <Accordion title="How is OrgX different from other AI tools?">
    Key differentiators: - **Human-in-the-loop**: Every agent action surfaces for
    your approval - **Evidence-based**: Artifacts include citations and
    verification proofs - **Optional integrations**: Add GitHub, Linear, and
    other supported connectors when your workflow needs them - **MCP-first**: Connect from ChatGPT,
    Claude, Cursor, and more - **Specialized agents**: Purpose-built domains
    (`engineering-agent`, `product-agent`, `marketing-agent`, `sales-agent`,
    `design-agent`, `operations-agent`, `orchestrator-agent`)
  </Accordion>

  <Accordion title="Is OrgX ready for production use?">
    OrgX is in **public beta**. Core features (web app, agents, decisions,
    hosted MCP, GitHub, and current Linear integration) are stable and used by
    production customers. Some features (Slack, Notion, enterprise SSO) are
    still in development. We recommend starting with supervised autonomy
    settings and gradually increasing trust.
  </Accordion>
</AccordionGroup>

## Agents

<AccordionGroup>
  <Accordion title="What can agents actually do?">
    Agents can:

    * **Read** from connected tools (issues, repos, documents)
    * **Analyze** and synthesize information
    * **Draft** specs, PRs, campaigns, reports
    * **Create** issues, PRs, comments in connected tools
    * **Propose** decisions for your approval

    Agents cannot:

    * Access tools you haven't connected
    * Ship work without your approval (unless you enable Full Auto mode)
    * Spend money without escalating
  </Accordion>

  <Accordion title="How do agents decide what to do?">
    When a task enters OrgX, the **orchestrator** analyzes:

    * Task title and description
    * Labels and metadata
    * Project context

    It then selects the best agent(s) and tools. You can see the reasoning in every decision card.
  </Accordion>

  {' '}

  <Accordion title="Can I customize agent behavior?">
    Yes, several ways: - **Autonomy levels**: Control how much agents can do
    independently - **Budget thresholds**: Set spending limits before escalation -
    **Scope limits**: Whitelist specific tools per agent - **Skill packs**: Custom
    instructions and capabilities (available in the OpenClaw plugin path;
    expanding across surfaces)
  </Accordion>

  <Accordion title="What if an agent makes a mistake?">
    Every agent output goes through:

    1. **Verification**: Automated checks for citations, policy, schema
    2. **Decision**: Human review before shipping

    If verification fails, you're notified to fix issues. If you catch something
    verification missed, reject the decision with feedback — the system learns.
  </Accordion>
</AccordionGroup>

## Integrations

<AccordionGroup>
  <Accordion title="Which tools does OrgX integrate with?">
    **Currently available**:

    * Linear (GA) — Full bidirectional sync
    * GitHub (Beta) — OAuth, PR creation
    * Stripe (GA) — Billing only

    **Coming soon**:

    * Slack — Notifications, blocker detection
    * Notion — Database sync
    * HubSpot — CRM integration
    * Figma — Design exports

    Some non-GA integrations expose MCP stub actions in selected surfaces.
    Exact availability can vary by workspace and client runtime.
  </Accordion>

  <Accordion title="Can I use OrgX without Linear?">
    Yes. Linear is optional. You can:

    * Start with the OrgX web app
    * Use the Wizard CLI or OpenClaw onboarding path
    * Use the ChatGPT or MCP interfaces
    * Connect GitHub only for code workflows

    If your team already uses Linear, connecting it gives OrgX richer issue and
    project sync, but it is not required to get value from the platform.
  </Accordion>

  <Accordion title="How do I disconnect an integration?">
    Go to **Settings → Integrations** and click **Disconnect** next to the
    integration. This revokes OrgX's access but doesn't delete historical data
    already synced.
  </Accordion>
</AccordionGroup>

## ChatGPT & MCP

<AccordionGroup>
  <Accordion title="How do I use OrgX in ChatGPT?">
    1. Enable **Developer Mode** in ChatGPT (**Settings → Apps & Connectors**)
    2. Add the OrgX MCP connector URL: `https://mcp.useorgx.com/mcp`
    3. Authenticate with your OrgX account when prompted
    4. Ask ChatGPT things like "Show my pending decisions" or "Approve the marketing campaign"

    See the [ChatGPT Integration guide](/docs/guides/chatgpt-integration) for details.
  </Accordion>

  <Accordion title="What is MCP?">
    The **Model Context Protocol** is an open standard for connecting AI assistants to external tools. OrgX implements MCP, which means you can connect from any MCP-compatible client—ChatGPT, Claude Desktop, Cursor, and others.

    See the [MCP Protocol reference](/docs/api/mcp-protocol) for technical details.
  </Accordion>

  <Accordion title="Can I use OrgX in Cursor?">
    Yes! Add OrgX to your `~/.cursor/mcp.json`:

    ```json theme={"dark"}
    {
      "mcpServers": {
        "orgx": {
          "command": "npx",
          "args": ["mcp-remote", "https://mcp.useorgx.com/mcp"]
        }
      }
    }
    ```

    OAuth sign-in happens automatically via browser on first connect. See the [MCP client setup guide](/docs/guides/cursor-mcp-setup) for full instructions across ChatGPT, Cursor, Claude, and VS Code.
  </Accordion>
</AccordionGroup>

## Pricing & Billing

<AccordionGroup>
  <Accordion title="How much does OrgX cost?">
    OrgX offers:

    * **Free tier**: 3 initiatives, 10 decisions/month, core OrgX workspace access
    * **Starter**: \$98/month for unlimited initiatives and 24/7 automation
    * **Team**: \$298/month for multiple seats and advanced integrations

    Enterprise pricing available on request. See [pricing](https://useorgx.com/pricing) for current details.
  </Accordion>

  <Accordion title="What counts against my usage?">
    Usage is based on:

    * Agent runs (number of tasks processed)
    * Token consumption (AI model usage)
    * Connected integrations

    Decision approvals and dashboard access are unlimited on all plans.
  </Accordion>

  <Accordion title="Can I cancel anytime?">
    Yes, subscriptions are month-to-month with no long-term commitment. Cancel
    in **Settings → Billing** and your access continues until the end of the
    billing period.
  </Accordion>
</AccordionGroup>

## Security & Privacy

<AccordionGroup>
  <Accordion title="Where is my data stored?">
    Data is stored in **Supabase Postgres** (US region by default, EU available
    on request). All data is encrypted at rest (AES-256) and in transit (TLS
    1.3).
  </Accordion>

  {' '}

  <Accordion title="Can agents see all my data?">
    Agents only access data you've explicitly granted via OAuth scopes. Each
    integration page shows exactly what permissions are requested. You can revoke
    access anytime.
  </Accordion>

  {' '}

  <Accordion title="Do you train AI models on my data?">
    No. Your data is used only to provide the OrgX service. We do not train
    foundation models on customer data. OpenAI calls are made with data usage
    disabled.
  </Accordion>

  <Accordion title="Is OrgX SOC 2 compliant?">
    We're currently preparing for SOC 2 Type I certification. See the [Security
    page](/docs/resources/security) for our current security posture and compliance
    roadmap.
  </Accordion>
</AccordionGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="OpenClaw pairing not working">
    Try:

    1. Make sure the OpenClaw browser extension is installed and enabled
    2. Sign out and sign back in to OrgX
    3. Click **"Pair with OpenClaw"** again from Settings
    4. Check the browser console for errors
    5. If using manual key fallback, verify the key in OpenClaw settings matches OrgX

    See the [OpenClaw setup guide](/docs/guides/openclaw-plugin-setup) for detailed troubleshooting.
  </Accordion>

  {' '}

  <Accordion title="Agents aren't doing anything">
    Check: 1. Is the agent running? Check status in Mission Control or via
    `orgx_spawn` 2. Are there pending decisions? Agents may be waiting for
    approval 3. Is the integration connected? Check Settings → Integrations 4. Are
    there errors? Check the agent's activity log
  </Accordion>

  {' '}

  <Accordion title="I'm not getting decision notifications">
    Check: 1. Notification settings in Settings → Notifications 2. Browser
    notification permissions 3. Email spam folder for digest emails 4. Decision
    urgency level (low urgency = badge only)
  </Accordion>

  {' '}

  <Accordion title="MCP connection failing">
    Try: 1. Disconnect and reconnect OrgX in your MCP client to refresh OAuth
    authorization 2. Check the endpoint URL is correct 3. Run health check: `npx
          mcp-remote ... --health-check` 4. Ensure network access to `mcp.useorgx.com`
  </Accordion>

  <Accordion title="Linear sync seems stuck">
    Try:

    1. Check webhook status in Settings → Integrations → Linear
    2. Trigger manual sync via the refresh button
    3. Disconnect and reconnect the integration
    4. Contact support if issues persist
  </Accordion>
</AccordionGroup>

## Still Have Questions?

<CardGroup cols={2}>
  <Card title="Contact Support" icon="envelope" href="mailto:hello@useorgx.com">
    Email us directly.
  </Card>

  <Card title="GitHub Issues" icon="github" href="https://github.com/useorgx/orgx-mcp/issues">
    Report docs, MCP, or integration issues.
  </Card>
</CardGroup>