Today's Pulse
Loading…
Total dev time
—
—
Commits
—
—
Context switches
—
today
Longest focus block
—
—
Time by project
Today
Focus timeline
— switches
Loading…
Toil detected
—
Loading…
Recent activity
Loading…
Context switches — past 7 days
High (>20)
Med
Low (<10)
Target: under 15 context switches/day for sustained deep work
This Week
—
Total this week
—
—
Commits
—
—
Avg focus block
—
this week
Best day
—
—
Daily hoursMon–Sun
Weekly breakdown
Time by project this week
Context switches — this weekAvg —/day
—
Activity log — this week
Loading…
Projects
Loading…
All projects breakdown
Project time share — this week
Most active branches
Loading…
Toil Detector
Repetitive command sequences detected in your shell history
Patterns found
—
detected
Total repetitions
—
all patterns
Est. time wasted
—
if not automated
Detected patterns—
Loading…
Applied automations0 applied
No automations applied yet.
Error fix knowledge
Learned fixes from your shell — match failing commands to what worked before
Embedding backend
—
—
Fix records
—
saved patterns
With vectors
—
semantic tier
Open windows
—
in progress
Open fix windowsneeds fix or fix-done
Loading…
Terminal commands
ghost fix-status
ghost fix-history
ghost fix-suggest "pytest tests/"
ghost fix-done
ghost fix-history
ghost fix-suggest "pytest tests/"
ghost fix-done
Look up a failing commandexact → fuzzy → semantic
Saved fixes—
| Loading… |
Focus Analysis
Context switching patterns and deep work insights
Avg focus block
—
this week
Best focus day
—
fewest switches
Worst day
—
most switches
Weekly switches
—
—
Focus block duration — this week
Bar height = longest focus block that day
Context switches — this week
Target: under 15/day. Red bars exceed threshold.
Activity heatmap — hourly output by dayPast 7 days
Intensity:
none
low
med
high
AI Insights
Natural language analysis of your workflow patterns
—
Workflow insights
Loading insights…
Ask Ghost PulseNatural language query
Configuration
ghost config · ~/.ghost-pulse/config.toml
LLM Provider
Provider
Which LLM backend to use
Model
Model name for the selected provider
—Ollama host
Base URL for local Ollama instance
—Collection
Shell hook
Capture every command via zsh/bash precmd
Git watcher
Track commits, branches, stash events
Window focus
Track active app for context switch detection
File watcher
Track .env and config file edits
Fix knowledge (RAG)
Auto-track fix windows
Open a window when a command fails; learn from the next steps
Embedding provider
For semantic search — auto tries local, Ollama, then OpenAI
Privacy
Local-only mode
Never send data to any external service
Retention
How long to keep event data
Web UI
Port
Port for ghost web server
8765Bind address
Network interface to bind to
127.0.0.1Daemon status
PID
—Uptime
—Events today
—DB size
—