GitHub hosts HunyuanVideo, Tencent's open-source framework for large-scale video generation models, enabling AI-driven video creation.
assaio
0Popularity
About assaio
assaio is an offline application that analyzes session logs generated by AI coding tools such as Claude Code, Codex, Gemini CLI, Copilot CLI, and Cline. It evaluates the effectiveness of AI-generated code by comparing output against tokens consumed, providing metrics on model fit, context health, adoption, and rework. The tool runs entirely locally without requiring an account, uploads, or network access, ensuring data privacy. It aggregates and normalizes data across multiple AI coding tools into a single report, allowing teams to assess which projects and tasks yield productive code versus those that result in excessive token consumption or circular reasoning. The analysis includes metrics such as AI lines produced per million tokens, edit loops, cache hygiene, and adoption trends across projects and tools. Reports are generated from data already present on the user’s machine, enabling comparisons over customizable time windows and task types.
Key features
- Offline analysis of AI coding tool logs
- Multi-tool support (Claude Code, Codex, Gemini CLI, Copilot CLI, Cline)
- Effectiveness measurement (AI lines per million tokens)
- Model fit analysis (premium vs cheaper model usage)
- Adoption tracking (sessions, active days, breadth)
- Edit loop detection (churn and rework)
- Cache hygiene assessment (context window usage)
- Customizable time windows and task-type filtering
Use cases
- Evaluating ROI of AI coding tool subscriptions by project
- Identifying debugging-heavy tasks that consume tokens inefficiently
- Comparing model performance across different coding tasks
Pros
- Analyzes session logs from multiple AI coding tools in one report
- Runs entirely offline with no data upload or account required
- Compares AI output against tokens consumed per project
- Supports macOS and Linux
- Provides detailed metrics on effectiveness, model fit, and adoption
Cons
- No Windows support
- No GUI interface
- Requires local session logs to function
Frequently asked questions about assaio
What does assaio do?
assaio analyzes session logs from AI coding tools like Claude Code, Codex, Gemini CLI, Copilot CLI, and Cline to evaluate the effectiveness of AI-generated code by comparing output against tokens consumed. It provides metrics on model fit, context health, adoption, and rework without requiring an account, uploads, or network access.
Who is assaio designed for?
The tool is designed for teams and organizations that use AI coding tools and want to assess the productivity and cost-effectiveness of their AI-generated code. It helps determine which projects and tasks yield productive code versus those that result in excessive token consumption or circular reasoning.
How does assaio work?
assaio reads session logs already present on the user’s machine and weighs the AI’s output against the tokens consumed per project. It aggregates and normalizes data across multiple AI coding tools into a single report, allowing comparisons over customizable time windows and task types.
Does assaio require an internet connection?
No, assaio runs entirely offline and does not require network access, uploads, or an account. It processes data locally to ensure data privacy.
What metrics does assaio provide?
assaio provides metrics such as AI lines produced per million tokens, edit loops, cache hygiene, adoption trends, model fit, context health, and rework. It also includes project-specific effectiveness ratios and session-level insights into token usage and output.
How do I get started with assaio?
To get started, install assaio via Homebrew on macOS or Linux using the command 'brew install assaio/tap/assaio-agent'. Run 'assaio-agent init' to detect your tools, review the directories it will read, and generate a report. A demo mode ('assaio-agent demo') is also available to preview reports using bundled sample data.
