Preview. These maps are our reading of public sources and haven't been checked with the projects' maintainers yet.

tare

A Claude Code skill from Kelviq that teaches Claude Code to read its own session logs and explain where your usage went and why you hit a limit.

Last push 2026-08-28, the day after its Show HN (88 points, 2026-08-27); 299 stars. Borderline: no commits for 59 days. (checked 2026-10-25)

In Coding agent usage trackers

Where it stands apart

• How is it mainly sustained? company with a paid product, tier, or ads (21% of 19 others)
• What does it mainly measure? where the tokens go (context size, per-turn cost, waste, files, tools) (14% of 21 others)
• Where do you mainly see it? inside the agent (status line, hook, plugin, mod, or skill) (14% of 21 others)

What it shares with most of the community

• Is it actively maintained? active (81% of 21 others)
• How is the main thing licensed? permissive open source (MIT/Apache/BSD) (62% of 21 others)
• Can it run entirely on your own device or infrastructure, without the maker's servers? yes (86% of 21 others)
• Do you need an account to use it? no account (71% of 21 others)
• Who is it for? developers (100% of 21 others)
• Does it only report, or does it change what the agent does? only reports and alerts; you decide (90% of 21 others)
• Does your usage data leave your machine? no; the only outside calls are to the AI provider itself or a public price list (71% of 21 others)
• What access to your provider accounts does it take? none; it reads local files or data the agent hands it (63% of 19 others)
• Whose usage is it built around? one developer's own usage (76% of 21 others)

Nearest neighbors

• Decant 87% of 15 shared answers match
• ClaudeStatsBar 86% of 14 shared answers match
• CostClaw 80% of 15 shared answers match

Every answer

Sources