OpenAI builds and deploys advanced AI models like GPT-4o for autonomous agents and workflows.
AGANAL
About AGANAL
AGANAL processes session logs generated by AI coding agents and converts them into metrics and visualizations. It reads local log files from supported tools such as Claude Code, OpenAI Codex, and Gemini CLI, then aggregates data like token usage, tool calls, context pressure, and retries into a unified dashboard. The tool provides three main views per session: an analysis dashboard with statistics and charts, a filterable event stream showing messages and tool interactions, and a raw JSONL view for inspecting underlying records. AGANAL operates entirely offline with no telemetry or account requirements, relying only on files already present on disk. It supports multiple agent formats and can be pointed at custom directories of session logs. A command-line interface allows further analysis and integration with other workflows.
Key features
- Reads session logs from disk
- Aggregates metrics per agent
- Dashboard with statistics and charts
- Filterable event stream
- Raw JSONL view
- Command-line interface
- Hand-off prompts for agents
- Supports multiple agent formats
Use cases
- Debugging agent behavior and errors
- Analyzing token usage and costs
- Reviewing tool call patterns across sessions
Pros
- Local-only processing with no data upload
- Supports multiple AI coding agents in one interface
- Provides three distinct views per session
- Open source and free to use
- No account or telemetry required
Cons
- macOS 14+ only
- Requires local log files from supported agents
- No native Windows or Linux support
Frequently asked questions about AGANAL
What is AGANAL and what does it do?
AGANAL is a local-only analytics tool for AI coding agents that reads session logs generated by tools like Claude Code, OpenAI Codex, and Gemini CLI. It converts these logs into metrics and visualizations, including token usage, tool calls, context pressure, and retries, without requiring an account or uploading data.
Which AI coding agents does AGANAL support?
AGANAL supports session logs from multiple agents, including Claude Code, OpenAI Codex, Gemini CLI, Qwen Code, Cursor, opencode, and Google Antigravity. It can also read logs from any custom directory of session files.
Does AGANAL require an internet connection or account to function?
No, AGANAL operates entirely offline and does not require an account. All analysis is performed locally on the user's machine, and no telemetry or data is uploaded.
What kind of views or outputs does AGANAL provide?
AGANAL offers three main views per session: an analysis dashboard with statistics and charts, a filterable event stream showing messages and tool interactions, and a raw JSONL view for inspecting underlying records.
How do I get started with AGANAL?
To get started, clone the AGANAL repository and run it using a single Swift command. The tool automatically discovers agent sessions or allows users to point it at a custom directory of logs.
Is AGANAL free and open source?
Yes, AGANAL is free and open source. Users can download the macOS app or view the source code on GitHub without any cost.