Skip to main content
Use this guide when you want an AI client to manage the Application Tracker for you. After the Reactive Resume MCP server is connected, the agent can use tools for the same application workflows available in the app: listing applications, creating records, importing rows, updating stages, adding notes, managing follow-ups, attaching documents, and running Application Copilot.

Prerequisites

1

Connect the MCP server

Follow Using the MCP Server to connect your MCP client with OAuth or an API key.
2

Confirm the application tools are visible

Ask your client to list the Reactive Resume tools. You should see tools such as list_applications, create_application, update_application, attach_application_document, and draft_application_message.
3

Decide how destructive actions should be handled

Tell your agent to ask before deleting applications, bulk-updating many records, or replacing attached documents.
Application Copilot tools require the same AI provider setup used by the app. Match scoring and resume tailoring work best when the application has a linked Reactive Resume and a job description.

Start with a pipeline review

Begin by asking the agent to inspect your current pipeline. This gives the agent valid application IDs and avoids duplicate records.
Useful review prompts:

Create applications

Use create_application when you already know the role details.
If you have a job posting, ask the agent to extract details first.
If the posting is private, paste the description into your prompt:

Import applications

Use import_applications when you already have spreadsheet rows. The Application Tracker accepts up to 500 imported rows at a time.

Update stages and notes

Use update_application for structured changes and add_application_note when you want an activity timeline entry without changing other fields.

Attach sent documents

Use document attachment when you want the tracker to store the exact resume PDF or cover-letter PDF you sent for an application.
attach_application_document accepts base64-encoded PDF bytes and contentType: "application/pdf". Some MCP clients hide that detail when they can read local files. If your client cannot read local files, upload the PDF from the web app instead.

Run Application Copilot

Application Copilot tools let an agent use the same AI workflows available in the application detail panel.
Review AI-generated resumes, cover letters, and messages before sending them. MCP tools can draft and save context, but you are responsible for the final content.

Maintain your pipeline

Ask the agent to do periodic cleanup with explicit confirmation before broad changes.

Prompt library

Use these prompts as starting points. Replace company names, roles, tags, dates, and file references with your own details.

Daily review

Research and capture

Applying

Interviews and follow-ups

Reporting

Suggested agent instruction

Add this to your MCP client’s project or session instructions when you want the agent to manage applications safely:

Troubleshooting