@hk_net/pi-timestamp
Pi extension that shows timestamps for user input and agent completion timing
Package details
Install @hk_net/pi-timestamp from npm and Pi will load the resources declared by the package manifest.
$ pi install npm:@hk_net/pi-timestamp- Package
@hk_net/pi-timestamp- Version
0.1.8- Published
- Jul 22, 2026
- Downloads
- 938/mo · 254/wk
- Author
- hk_net
- License
- EUPL-1.2
- Types
- extension
- Size
- 24.9 KB
- Dependencies
- 0 dependencies · 2 peers
Pi manifest JSON
{
"extensions": [
"./timestamp.ts"
]
}Security note
Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.
README
pi-timestamp
Shows timestamps for user input and agent completion timing. Requires pi >=0.81.1.
What it does
- User input: Shows
Sent HH:MM:SSas a dim status line in the chat display after each user message - Agent completion: Shows
Done at HH:MM:SS · durationas a dim status line in the chat display after each agent turn (e.g.,Done at 14:32:05 · 3.2s) - Session/runtime summaries: Shows accent-colored summaries when switching sessions and at final Pi exit, including start/end times and durations. The final summary includes the complete Pi process runtime and every session interval.
All timestamps and summaries are display-only — session-switch summaries render in Pi's UI and the final runtime summary prints after Pi restores the terminal in TUI mode. They never enter the LLM context.
Display behavior
Timestamps appear inline in the chat display, similar to Pi's built-in tool timing lines such as Took 1.9s. They are not appended to the session as user or custom messages, so they do not pollute the model context.
Installation
pi install npm:@hk_net/pi-timestamp
Or copy manually:
mkdir -p ~/.pi/agent/extensions
cp packages/pi-timestamp/timestamp.ts ~/.pi/agent/extensions/timestamp.ts
Duration measurement
Captured from agent_start (when the model starts processing) to agent_end (when all tool calls and model processing are done).
Issues and feedback
Found a bug or have a feature request? Please report it on GitHub Issues.
For security vulnerabilities, please use GitHub's private vulnerability reporting instead of opening a public issue.