Codag provides a log compression service that transforms massive system logs into compact, token‑efficient text, preserving the exact lines and context needed for debugging. By filtering out routine noise and ranking suspect patterns, it reduces token usage by up to 95% while delivering real line numbers, enabling AI agents like Claude or GPT to process logs quickly and accurately.
Funding
Funding not disclosed
Founders
Product
Problem
Developers and operations teams need to feed system logs into AI agents for debugging, but raw logs contain millions of lines and consume excessive token counts, often exceeding the context limits of language models. This makes it costly or impossible for agents to process logs effectively.
Solution
Codag offers a log compression service that analyzes raw log streams, extracts the most relevant patterns, and returns compact text containing only the lines that matter, each with real line references and nearby context. By ranking patterns by signal strength and preserving exact log lines, the service reduces token counts by up to 95% while maintaining the full debugging signal. The output is plain text that can be pasted directly into any LLM or debugging tool without additional transformation. Integration is achieved via a simple CLI wrapper that can be applied to any existing log command, enabling seamless adoption across environments. A free tier provides unlimited usage for developers, allowing them to test and use the service without cost.
Target Audience
Primary users are developers, SREs, and platform engineers who need to provide log data to AI debugging agents, as well as teams building AI‑assisted monitoring or incident‑response tools.
Features
- Automatic extraction and ranking of high‑signal log patterns with real line numbers
- Token reduction of up to 95% (e.g., 1.2 M lines compressed to ~3.3 k tokens)
- Plain‑text compact output that can be fed to Claude, GPT, or other agents
- Drop‑in CLI wrapper supporting common log commands (kubectl, docker, aws, vercel, etc.)
- HTTPS API for programmatic access and integration with existing pipelines
- Free tier with unlimited usage for developers