Porch

Product Introduction

1. Definition: Porch is a local, open-source desktop application for macOS that functions as an activity tracker and analytics dashboard for developers using Anthropic's Claude Code and Claude Codex AI coding assistants. It operates as a system-level observer, aggregating session data from across terminals and code editors.
2. Core Value Proposition: Porch exists to solve developer visibility and productivity tracking challenges when using AI coding tools. It automatically transforms fragmented, ephemeral Claude interactions into structured, time-based summaries and actionable insights, enabling better project management, cost tracking, and workflow optimization without requiring a cloud account.

Main Features

1. Cross-Editor Session Aggregation: Porch monitors Claude Code and Codex activity system-wide, capturing sessions from any terminal (iTerm2, Warp, macOS Terminal) or supported code editor (like VS Code). It uses low-level system monitoring to unify these discrete interactions into a single timeline per project, providing a holistic view of AI-assisted development work.
2. Automated Time & Project Summarization: The core engine parses raw session logs to generate daily, weekly, and monthly summaries organized by project directory. It employs heuristic analysis to estimate task duration and identify potential "blockers"—periods of high back-and-forth interaction or repeated code generation—flagging them for review. Summaries are exported as clean Markdown.
3. Local-First Analytics & Cost Tracking: All session data, status, and usage metrics are stored exclusively on the user's Mac. Porch calculates token usage per session and provides API-cost equivalency estimates based on current Anthropic API pricing, allowing developers to monitor and forecast expenses without sending data to a third party. The open-source nature allows for verification and customization of these calculations.

Problems Solved

1. Pain Point: Developers lack a coherent audit trail of their AI coding assistant usage, leading to fragmented context, difficulty in estimating time spent on AI-aided tasks, and no clear way to track cumulative token consumption or its associated costs.
2. Target Audience: Primary users are software engineers, DevOps professionals, and technical leads who regularly use Claude Code/Codex within their local development environment on macOS. Secondary users include engineering managers seeking to understand team AI tool usage patterns and costs.
3. Use Cases: Essential for freelancers tracking billable hours enhanced by AI, teams auditing AI tool spending, developers conducting retrospectives on problem-solving sessions, and individuals building a personal knowledge base in tools like Obsidian from their coding workflow.

Unique Advantages

1. Strengths & Limitations (Pros & Cons):

   • Pros: Unmatched privacy and data sovereignty (everything stays local). Transparent and customizable (open-source). Provides unique "blocker" detection for workflow optimization. Direct integration with local note-taking apps via Markdown export. No recurring subscription fee.
   • Cons: Platform-locked to macOS. Requires user to have Claude Code/Codex installed and signed-in. Analysis is retrospective, not real-time. Heuristic time/blocker estimates may not be perfectly accurate. Lacks cloud-based collaboration or team dashboard features.

2. Key Alternatives & Differentiation:

   • Manual Logging / Spreadsheets: The default alternative is manual tracking. Porch differentiates by providing fully automated, passive data collection and structured analysis, saving significant time and increasing accuracy.
   • Cloud-Based AI Analytics Platforms (e.g., tools for GitHub Copilot): These often require sending code snippets or metadata to the cloud. Porch's fundamental differentiation is its strict local-first, privacy-centric architecture, appealing to security-conscious developers and organizations.
   • Anthropic API Console: Provides raw token counts for direct API calls but not for the Claude Code desktop application. Porch fills this gap by estimating costs for the desktop app experience, which operates on a different billing model.

Frequently Asked Questions (FAQ)

1. Does Porch send my code or session data to the cloud? No. Porch is architected on a local-first principle. All session monitoring, processing, and summary generation occurs directly on your Mac. Your code and interaction data never leave your machine unless you explicitly export a summary document.
2. How does Porch estimate task time and API costs for Claude Code? Porch analyzes session logs, measuring activity duration and calculating token usage from the interactions. It then applies the public pricing rates of the Anthropic API to these token counts to generate a cost equivalency estimate, as the desktop app itself uses a subscription model.
3. Can I use Porch with other AI coding assistants like GitHub Copilot or Cursor? No. Porch is specifically engineered to parse and interpret the activity logs and data structures generated by Anthropic's Claude Code and Claude Codex applications. It is not compatible with assistants from other providers.
4. Is Porch free to use? Yes. Porch is released as open-source software, which means it is free to download, use, and modify. There is no requirement for a Porch-specific account or subscription.
5. How do I export summaries to Obsidian? Porch saves summaries as standard Markdown (.md) files. You can configure Porch to export these files directly into your Obsidian vault folder or any other designated directory on your system. Obsidian will then automatically detect and index them.