You can now bring the context of your contracting business into an AI assistant through the Plan@Job MCP server. Connect a compatible AI app to your own Plan@Job account, review the permissions, then ask for help with account setup, existing jobs and internal follow-ups.
The connection uses Model Context Protocol (MCP), a standard that lets AI applications discover and use tools from another service. For a contractor, the useful part is simple: your assistant can work with permitted information from your business instead of relying on a job description you copied into a chat last week.
Start with the Plan@Job AI connection guide. The remote MCP server URL is:
https://planajob.com/api/mcp
This is a manual custom connection. Provider account, region and workspace restrictions still apply. The instructions below describe the providers' documented setup routes; they are not a claim of marketplace approval or a completed Plan@Job connection test in every app.
What can the Plan@Job MCP server do?
The connector currently provides 16 business tools and a saved-action receipt lookup. They cover practical parts of running a contracting business:
- Check account readiness. Find missing company information and see what needs completing next.
- Improve your profile. Update supported personal details, company information and work preferences within your existing permissions.
- Review existing work. Read your company's private or won jobs and inspect the subcontracting desk.
- Prepare the team. Create an internal subcontracting workspace for an eligible job and save its brief.
- Organise follow-ups. Review saved contact history, create an internal follow-up, assign it or change its due date.
- Check a saved action. Retrieve the action receipt when a connection was interrupted or you need to confirm what happened.
For example, try: "Review my Plan@Job jobs and tell me which ones need attention. Explain the next useful action for each."
For account setup, try: "Check my Plan@Job company profile. Suggest improvements using only the facts already saved and the details I give you. Show me the proposed changes before saving."
The connector does not send emails, place calls, award work, make payments or change bank details. Saving an internal follow-up does not contact the subcontractor. These boundaries keep the distinction between preparing work and committing to an external action clear.
Before you connect
You need a Plan@Job contractor account and an AI app that supports remote MCP with OAuth sign-in. Account setup tools are available before onboarding is complete; operational tools retain the account's normal onboarding restrictions.
Open Settings > Connect your AI in Plan@Job, or visit the public setup guide. Copy the connection URL. Add it in your AI app, then sign in on Plan@Job and check the company, requesting app, return address and requested permissions.
You choose whether to grant account reading, account updates, work reading or internal work updates. Start with the permissions needed for your task. Never paste your Plan@Job password or a service key into a chat.
How to connect Plan@Job to ChatGPT
Where your account or workspace permits custom connections:
- Open ChatGPT settings and enable Developer mode under Security and login.
- Open ChatGPT Plugins, select +, and add a connection named Plan@Job.
- Enter a short description and
https://planajob.com/api/mcpas the MCP server URL. - Complete OAuth sign-in on Plan@Job and choose the access to grant.
- Enable the connection in a new conversation and start with an account-readiness question.
ChatGPT's menus and workspace permissions can change. If the option is missing, check the official ChatGPT connection instructions and your workspace's settings.
How to connect Plan@Job to Claude
- Open Customize > Connectors in Claude.
- Select + > Add custom connector.
- Name it Plan@Job and enter
https://planajob.com/api/mcp. - Follow the sign-in flow and approve your chosen Plan@Job permissions.
- Enable the connector for the conversation and ask it to check your account setup.
Team and Enterprise workspaces may require an owner to add the connector for the organisation first. Follow Claude's official remote MCP guide for your plan and device.
How to connect Plan@Job to Grok
Grok documents custom MCP connectors in its web app:
- Open Grok Connectors.
- Choose New Connector > Custom.
- Enter the Plan@Job MCP URL and complete authentication.
- Check the company and permissions on the Plan@Job consent screen before allowing access.
- Start with a read-only question and verify the result against your Plan@Job workspace.
Business and Enterprise users may need their organisation administrator to provision the connection. See the official Grok connector guide.
Can I use Plan@Job with Gemini?
There are different Gemini products, and their connection options are different.
Gemini Spark: Google documents custom MCP apps for Spark tasks. Its stated eligibility includes an adult personal-account user in the United States, English and Keep Activity enabled. This is not a general promise of custom connectors in ordinary Gemini chat or availability in the UK. Eligible users can find the option under Settings & help > Connected Apps > Custom apps for Spark. Check Google's setup and eligibility guidance before starting.
Gemini CLI: Google's command-line assistant supports remote HTTP MCP servers and browser-based OAuth. For users who already have Gemini CLI installed, add the server:
gemini mcp add --transport http planajob https://planajob.com/api/mcp
Then run /mcp auth planajob inside Gemini CLI and complete the Plan@Job sign-in flow. Use /mcp list to check the connection. The lowercase planajob is the local connection identifier; the product name is Plan@Job. See the Gemini CLI MCP guide for current instructions.
What about other AI platforms?
The same remote MCP endpoint can be used as the basis for integrations with other compatible clients. It does not require a separate Plan@Job database for each AI provider.
Microsoft documents adding an existing MCP server to a Copilot Studio agent, including OAuth configuration. This is a route for a configured business agent, rather than a promise that every consumer Copilot account accepts a custom connector. See Microsoft's MCP setup guide.
Developers can also build integrations using the Gemini API's remote MCP support or xAI's remote MCP tools. The application must handle the user's authentication, token refresh and permitted actions; an API integration is not the same as installing a connector in a consumer chat app.
Sintra: We have not verified a documented route for installing an arbitrary custom MCP server in Sintra. Its integration documentation describes built-in integrations. Plan@Job is therefore not advertised here as an available Sintra integration.
Your company and permissions stay in control
The MCP connection uses your existing company membership and role. It does not give an AI access to platform administration, source code or the whole database. Company profile changes still require the appropriate company permissions.
Supported writes require write access and produce saved-action receipts. If an action's completion is uncertain, check its receipt and the affected record before asking for the action again.
You can disconnect from Settings > Connect your AI. Disconnecting stops future access checks; an action already in progress may finish. It does not remove information already received by your AI provider, which remains subject to that provider's privacy and retention settings.
Frequently asked questions
Do I need to install software on the Plan@Job server?
No. Plan@Job hosts the remote MCP endpoint. You add its URL to a compatible AI app and authorise your account. Gemini CLI users separately need the CLI installed on their own computer.
Is this the same as Ask Plan@Job?
No. Ask Plan@Job is the assistant inside the platform. The MCP connector brings supported Plan@Job tools into an external AI app that you choose.
Does connecting an AI bypass onboarding or vetting?
No. Account setup remains available, but operational tools keep the account's normal restrictions. The connector does not change the quotation or vetting process.
Is Plan@Job listed in every AI marketplace?
No. A working remote MCP endpoint and a provider's custom-connector feature are separate from marketplace submission and approval. Use the manual setup routes above where your account supports them.
What should I try first?
Ask: "Check my Plan@Job account readiness and list the information I should complete next." Then compare the response with your account before granting any additional permissions.
Ready to begin? Open the Plan@Job MCP installation guide, copy the URL and connect the AI assistant you already use. If you are new to the platform, explore Plan@Job for contractors or create your account.