AI-powered website builder that generates a professional site in under a minute and includes tools for content, SEO, and analytics.
Rustly
Local observability tool for coding agents that tracks token usage and context growth
About Rustly
Rustly is a local observability tool designed to track and analyze how coding agents consume tokens, context windows, and API quotas. It sits between coding agents like Claude Code, Codex, or Cursor and their respective providers, recording every request without altering the workflow. The tool provides detailed breakdowns of token expenditure, cache behavior, tool schema usage, and context growth on a per-turn basis. Rustly operates entirely offline, storing all data in a local SQLite file and never transmitting information outside the user’s machine. It includes an advice engine that generates actionable recommendations based on observed usage patterns, ranked by potential savings. The tool also offers enforcement rules that can modify tool calls to reduce unnecessary token consumption, such as limiting file reads or skipping redundant operations. Historical data from past sessions is automatically imported, allowing users to analyze months of agent activity immediately after installation.
Key features
- Token and context window tracking per turn
- Cache behavior analysis and breakdown
- Tool schema usage auditing
- Cost tracking with API-equivalent values
- Advice engine with step-by-step instructions
- Enforcement rules for tool call modifications
- Historical data import from agent transcripts
- Local SQLite storage for all data
Use cases
- Debugging excessive token consumption in coding agents
- Optimizing agent configurations to reduce costs
- Identifying redundant or unnecessary tool calls
Pros
- Records every request without sampling or estimation
- Operates entirely offline with no telemetry or cloud storage
- Provides per-turn breakdown of token usage and context growth
- Includes an advice engine that generates ranked recommendations
- Supports enforcement rules to modify tool calls and reduce costs
Cons
- No free tier after 14-day trial
- No cloud-based version or multi-user support
- Limited to macOS and Linux platforms
- Requires local model for advice generation
Frequently asked questions about Rustly
What is Rustly and what does it do?
Rustly is a local observability tool that tracks and analyzes how coding agents consume tokens, context windows, and API quotas. It records every request made by agents like Claude Code, Codex, or Cursor without altering the workflow, providing detailed breakdowns of token expenditure and usage patterns.
Who is Rustly designed for?
Rustly is designed for developers and teams using coding agents such as Claude Code, Codex, Cursor, or similar tools. It is particularly useful for those looking to optimize token usage, reduce costs, and improve the efficiency of their agent-based workflows.
How does Rustly handle privacy and data storage?
Rustly operates entirely offline, storing all data in a local SQLite file on the user’s machine. It never transmits information outside the user’s device, ensuring complete privacy and control over sensitive data.
Can Rustly enforce rules to reduce token consumption?
Yes, Rustly includes enforcement rules that can modify tool calls to reduce unnecessary token consumption, such as limiting file reads or skipping redundant operations. These rules act through hooks that the agents respect, providing transparency and control.
Does Rustly provide historical data analysis?
Yes, Rustly automatically imports historical data from past sessions, allowing users to analyze months of agent activity immediately after installation. This includes data from tools like Claude Code and Codex, which store full transcripts on disk.
What is the pricing model for Rustly?
Rustly is available as a one-time purchase for $99, with no subscription or usage fees. A free trial is available for 14 days without requiring a credit card or account.