Ghosted

Overview

Ghosted is a terminal-based job application tracker designed for developers navigating the modern hiring process. Built in collaboration with Claude Code, it provides both an interactive TUI for daily use and a CLI for automation—making it ideal for integration with AI agents that can parse job postings and update statuses automatically.

The project is open-sourced under the MIT license.

Dual Interface Design

The project offers two ways to interact: a rich Terminal User Interface for exploration and manual updates, and a CLI for programmatic access. Both share the same JSON data store, so changes sync instantly.

The CLI was specifically designed for AI agent integration—commands accept and output JSON, making it trivial to build automation that parses job postings from emails or websites and automatically creates tracked applications.

Status Pipeline

Applications flow through an 8-stage pipeline reflecting the real recruitment journey:

1. Saved - Bookmarked for later
2. Applied - Application submitted
3. Screening - Initial review
4. Interview - In progress
5. Offer - Received
6. Accepted - Job secured
7. Rejected - Application declined
8. Withdrawn - Candidate withdrew

Quick keyboard shortcuts (1-8) let you update status without leaving the list view.

Rich Data Tracking

Each application tracks comprehensive details beyond just company and status:

• Compensation: Salary range with smart formatting ($150k display)
• Contacts: Recruiter name and email for follow-ups
• Documents: Links to resume version and cover letter used
• Interviews: Multiple interviews with type, date, interviewer, and notes
• Follow-ups: Next action reminders

Technical Highlights

• Bubble Tea Framework: Elm-inspired architecture with clean state management and a view state machine
• Zero Database Overhead: Plain JSON files make data portable, version-controllable, and human-readable
• Smart Partial Matching: CLI accepts partial UUIDs (ghosted get abc finds first match starting with “abc”)
• Cross-Platform: Single Go binary compiles for macOS, Linux, and Windows
• Sample Data Seeding: New installations auto-populate with example applications for immediate exploration

Building with Claude Code

This project was developed in collaboration with Claude Code, Anthropic’s AI coding assistant. The pairing workflow proved especially effective for:

• Designing the data model and status state machine
• Implementing the Bubble Tea view architecture
• Building the CLI command structure with proper flag handling
• Writing comprehensive documentation

The experience reinforced how AI-assisted development can accelerate solo projects while maintaining code quality and architectural consistency.

Roadmap

Looking ahead, Ghosted has room to grow into a more comprehensive job search companion. Here are the planned expansions:

Embedded Agent Integration

The current CLI-first design enables external AI agents to interact with Ghosted, but the next step is bringing the agent inside the tool itself. An embedded agent could monitor your application pipeline, suggest when to follow up on stale applications, draft outreach messages, and even analyze job posting patterns to identify which opportunities match your profile best. The goal is proactive assistance without leaving the terminal.

Add Postings via Clipboard or Drag-and-Drop

Currently, adding job postings requires manual entry or external automation. A more fluid workflow would let you paste a job posting URL or description directly from your clipboard, or drag-and-drop a file path into the TUI. The system would parse key details—company name, role title, requirements, compensation—and pre-populate the application record, reducing friction between finding a job and tracking it.

Add Resumes via Clipboard or Drag-and-Drop

Tracking which resume version you sent to each company is easy to forget. This feature would let you attach resume files to applications by pasting a path or dragging the file into the interface. Each application would link to the specific resume used, making it simple to reference what the employer has seen and tailor your interview prep accordingly.

Add Cover Letters via Clipboard or Drag-and-Drop

Similarly, cover letters deserve the same treatment. Drag-and-drop or clipboard input would associate cover letter files with their applications, building a complete document trail. When you land an interview, you’d have instant access to exactly what you wrote—no more hunting through folders trying to match files to companies.