Product Introduction
Definition: Claude Code & Codex Usage Trading Cards by Rudel is a sophisticated developer analytics and data visualization platform specifically designed for AI-augmented programming environments. It functions as a "Developer Wrapped" service that ingests telemetry from Claude Code and Codex sessions to generate procedurally derived digital trading cards. These cards categorize developer behavior into specific "archetypes" using multi-dimensional data points.
Core Value Proposition: The product bridges the gap between raw LLM (Large Language Model) usage logs and actionable developer insights. By quantifying session shapes, token consumption, and model distribution, it provides developers with a gamified yet highly technical overview of their AI interaction efficiency. It serves as a diagnostic tool for optimizing AI spend, identifying skill gaps in prompt engineering, and visualizing the ROI of agentic workflows.
Main Features
Automated Archetype Classification: The system utilizes a proprietary algorithm to analyze session patterns, including "session shape" (intensity over time), error-to-success ratios, and command diversity. Based on these heuristics, it assigns a specific persona—such as the "Maniac" for high-intensity, high-frequency users—providing a high-level summary of the developer's operational style within AI coding interfaces.
Multi-Model Telemetry & Token Analysis: Rudel tracks granular usage metrics across different AI backends, specifically Claude (Anthropic) and Codex (OpenAI) models. It parses input/output token ratios (e.g., 1.2M input vs. 740K output) to determine the "Model Mix." This feature allows users to visualize their reliance on specific LLMs and understand the cost-to-output dynamics of their 1.9M+ total token transactions.
Agentic Workflow & Skill Tracking: The platform monitors the usage of "sub-agents" and specific "skills" (such as Refactor, Search, or Test) invoked during a session. By logging command frequency and repository breadth (the number of different codebases touched), it provides a heatmap of technical versatility and the complexity of tasks delegated to AI agents.
Financial & Efficiency Auditing: Rudel calculates high-level economic metrics, including total expenditure (e.g., $347) and the "Dollar per Commit" ratio ($3.30 in sample data). By correlating "Commit Rate %" (e.g., 48%) with "Success Rate %" (e.g., 69%), it offers a quantitative measure of how effectively the AI is contributing to the version control history of a project.
Problems Solved
Lack of AI Spend Visibility: Many developers use Claude Code or Codex without a clear understanding of their accumulated costs or token efficiency. Rudel solves this by centralizing spend data and calculating the cost-effectiveness of AI-generated commits.
Opaque Developer Patterns: It is difficult for engineers to objectively evaluate how they interact with AI—whether they are over-relying on refactoring or failing to utilize sub-agents. Rudel’s "archetype" system and "Fav Skill" identifier provide immediate feedback on these behavioral patterns.
Target Audience: This tool is essential for Full-stack Developers, AI Engineers, and DevOps professionals who utilize terminal-based AI coding assistants. It is also highly valuable for Engineering Managers looking to understand the productivity gains and usage habits of their team in an AI-first development environment.
Use Cases:
- Year-end productivity reviews for individual contributors.
- Optimizing AI usage by identifying high-error session patterns.
- Sharing development milestones within the open-source community via social-friendly trading cards.
- Self-hosting usage analytics for privacy-conscious enterprise environments.
Unique Advantages
Open Source & Self-Hostable: Unlike proprietary analytics dashboards, Rudel is open-source. This allows developers to maintain full control over their session data and telemetry logs, ensuring that sensitive IP within their prompts or repository names remains secure within their own infrastructure.
Deep Contextual Integration: While standard dashboards show simple API calls, Rudel analyzes the "shape" of the session. It looks at repo breadth and the specific sub-agents used, providing a 360-degree view of the "agentic" development process rather than just raw character counts.
Gamified Performance Metrics: By converting dry technical data into "Trading Cards," the product increases engagement with performance metrics. This unique approach turns routine auditing into a shareable asset, fostering a culture of transparency and continuous improvement in AI prompt engineering.
Frequently Asked Questions (FAQ)
How does Rudel determine my Claude Code "Archetype"? The archetype is determined by analyzing a combination of your session intensity, the variety of commands used (e.g., shell vs. refactor), your success-to-error ratio, and the breadth of repositories you interact with. For instance, a "Maniac" archetype typically reflects high session frequency with a high volume of input/output tokens over a short period.
Can I use the Rudel Trading Card generator for free? Yes, the product is free and open-source. Users can access the hosted version at app.rudel.ai/wrapped or choose to self-host the application to ensure their coding telemetry remains private and within their local environment.
What models are supported for the "Model Mix" visualization? Rudel currently tracks usage across Claude (Anthropic) and Codex (OpenAI) environments. It distinguishes between the input and output tokens for these models to help you understand which AI provider is driving your most successful code commits and where your budget is being allocated.
