- Home
- Skills
- Productivity
- tps-meter Setup: TTFT & Tokens/s Status Line
tps-meter Setup: TTFT & Tokens/s Status Line
A guided, confirm-before-you-run setup for tps-meter, the free open-source plugin that shows time to first token and output tokens/s in Claude Code.
Free
Works with the AI Agents you already use
About this skill
tps-meter-setup is an installer-style Agent Skill. It walks your coding agent through setting up tps-meter, a plugin for Claude Code that measures how fast the model answers: time to first token (TTFT) and output tokens per second for every main-conversation request. The numbers are written to a small JSON file, and the skill helps you show them in your status line, for example:
Opus | ⚡ 121 tok/s · TTFT 1.1s · avg 129 tok/s
What the skill does
- Checks your Claude Code version and explains the risk: tps-meter uses an early-access function hook API and was tested by its author on Claude Code 2.1.296.
- Installs the plugin with
claude plugin marketplace add kadaliao/claude-modsandclaude plugin install tps-meter@kadaliao. - Explains that metrics are written to
~/.claude/tps/<session_id>.json(or$CLAUDE_CONFIG_DIR/tps/). - Optionally installs an example status line script (requires
jq) and merges thestatusLinesetting without clobbering your existing config. - Verifies the file is being written, and covers update, uninstall and troubleshooting.
Safety by design
- The agent is instructed to tell you about every network access (GitHub) and every path it will write, and to get your confirmation before each step.
- It never overwrites an existing status line or settings entry without showing it to you first.
- No secrets, no telemetry, no binaries: the package is plain text (Markdown plus one shell script).
Good to know
- Early-access API: it may change between Claude Code releases. If it breaks, the status line simply shows nothing extra.
- Only the main conversation is measured, not subagents.
- Very short requests (under 50 tokens or 250 ms of streaming) produce no per-request rate; the last usable value is kept.
What is included
SKILL.md: the step-by-step instructionsreferences/metrics.md: what every metric and JSON field meansexamples/statusline-command.sh: example status line scriptREADME.md
License and source: MIT, Copyright (c) 2026 Kada Liao. Source: https://github.com/kadaliao/claude-mods
Early-access API, may break on Claude Code updates. Measures only the main conversation, not subagents. Very short requests (under 50 tokens or 250 ms) produce no per-request rate.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
2 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 27 days ago
- Free with an Agensi account
Requires
Permissions
Network access is limited to GitHub (kadaliao/claude-mods) for plugin install; the skill instructs the agent to ask for confirmation before each step. Writes only under ~/.claude/.