Developer Creates ESP32 Desktop Dashboard for Claude Code Usage Monitoring

Hermann Bjorgvin has built Clawdmeter, a physical desktop dashboard that displays Claude Code usage statistics on a compact ESP32-based device. The project utilizes a Waveshare ESP32-S3-Touch-AMOLED-2.16 to present token consumption data on a 2.16-inch AMOLED screen, creating a persistent visual interface for developers tracking their AI coding assistant usage.

Hardware Foundation

The Clawdmeter implementation centers on Waveshare's ESP32-S3-Touch-AMOLED-2.16 development board, which combines the ESP32-S3 microcontroller with a high-contrast AMOLED display. This particular board configuration provides the processing capability for data retrieval and network connectivity alongside a screen resolution suitable for displaying numerical usage metrics without requiring external display drivers or complex interfacing.

The ESP32-S3 variant offers dual-core processing at up to 240MHz, integrated Wi-Fi connectivity, and sufficient GPIO for future expansion while maintaining the low power profile characteristic of the ESP32 family. The touch-enabled AMOLED display eliminates the need for physical controls while providing clear readability under varying ambient lighting conditions typical of developer workspaces.

Data Source Integration

Clawdmeter accesses usage information from the local filesystem location where Claude Code stores operational data. The application reads from the ~/.claude/ directory, which contains usage logs and token consumption records maintained by Anthropic's local coding assistant client.

This approach leverages the existing data persistence mechanisms within Claude Code rather than requiring API calls or external service dependencies. The dashboard reads locally stored usage files, parsing token consumption metrics that Claude Code automatically tracks during development sessions. This design choice ensures the monitoring system operates independently of network connectivity once initial configuration is complete.

Display Implementation

The 2.16-inch AMOLED screen presents token usage information in a format optimized for at-a-glance consumption. AMOLED technology provides high contrast ratios and power efficiency when displaying predominantly black backgrounds with bright text elements, making it well-suited for numerical dashboard applications that update periodically rather than requiring constant screen refresh.

The display dimensions accommodate multiple data points simultaneously while maintaining readability from typical desk viewing distances. Token usage statistics can include current session consumption, daily totals, and potentially historical trends depending on the data available in the Claude Code storage directory.

Developer Workflow Integration

Physical usage displays represent a departure from software-based monitoring approaches that require active window switching or notification systems. By presenting usage metrics on dedicated hardware, developers maintain awareness of their AI assistant consumption without interrupting their primary coding workflows.

This mirrors patterns I observed during the early cloud computing transition in the 2010s, when developers similarly sought persistent visibility into resource consumption that was previously invisible or abstracted. Physical dashboards for digital metrics served as training wheels for understanding new cost models, whether AWS billing or, in this case, token-based AI usage.

The persistent visibility approach addresses a common challenge with usage-based AI services, where consumption can accumulate rapidly during intensive coding sessions without clear feedback mechanisms. Traditional software notifications compete for attention with other development tools, while a dedicated physical display provides continuous peripheral awareness.

Technical Architecture Considerations

The project's reliance on local file system access suggests a polling-based architecture where the ESP32 periodically reads updated usage data from the ~/.claude/ directory. This requires file sharing or network access from the ESP32 to the development machine running Claude Code, likely implemented through SMB, NFS, or HTTP-based file serving.

Network connectivity requirements depend on whether the ESP32 runs a web client polling a local server or accesses shared storage directly. Either approach introduces considerations around authentication, network reliability, and data synchronization timing that affect the accuracy and timeliness of displayed metrics.

The ESP32-S3's dual-core architecture enables separation of network operations from display management, ensuring UI responsiveness even during data retrieval operations. Touch functionality could support multiple display modes or configuration options without requiring physical controls.

Broader Context and Implications

Physical dashboards for digital services represent a growing category as cloud-native and AI-based development tools increasingly abstract resource consumption from direct user visibility. The Clawdmeter approach provides a template for other usage monitoring applications where developers benefit from persistent awareness of consumption-based service usage.

The open-source nature of the project, available through GitHub, enables customization for different AI services or usage tracking requirements. The hardware platform supports modification for monitoring multiple services simultaneously or integrating additional metrics beyond token usage.

As AI coding assistants become standard development tools, usage awareness mechanisms will likely evolve from novelty projects like Clawdmeter toward integrated features within development environments. However, the dedicated hardware approach offers advantages in terms of visibility and workflow integration that software-only solutions struggle to match.

The project demonstrates the continued relevance of purpose-built hardware solutions even as software capabilities expand, particularly for applications requiring persistent visibility or reliable operation independent of primary development system resources.