Track how you actually spend your time as a developer. AI sessions · GitHub commits · Jira tickets · Tempo worklogs — all in one dark-themed dashboard.
devtrace parses your local Claude Code and Codex CLI session logs, fetches GitHub commits via the gh CLI, and renders everything in a Streamlit dashboard with daily, weekly, and monthly breakdowns.
It answers questions like:
- How many hours did I actually code this week?
- Which projects are eating my time?
- When during the day am I most productive?
- How many tokens am I burning through Claude?
- Does my logged Jira time match what I actually worked?
Hours, sessions, and commits broken down by day. Filter by date range, source (Claude / Codex / GitHub), and project.
See exactly when during the day your sessions happened — and a heatmap showing your coding patterns across the week.
The files you change most, ranked by lines added + removed. Drill down by project.
Input / output / cache tokens broken down by model, project, and day. Track your AI spend over time.
Connect your GitHub account, select which orgs and repos to track, and set up project-to-folder mappings.
Jira tickets with estimated vs. logged hours — see where estimates diverge from reality.
Match your AI sessions to Jira tickets and bulk-submit Tempo worklogs without leaving the dashboard.
git clone https://github.com/yakupkeskin/devtrace.git
cd devtrace
uv sync
uv run streamlit run devtrace/dashboard/app.pyOpen http://localhost:8501 and configure your GitHub account from the GitHub settings tab in the top navbar.
docker run -p 8501:8501 \
-v ~/.claude/projects:/data/claude:ro \
-v ~/.codex/sessions:/data/codex:ro \
-v ~/.config/gh:/root/.config/gh:ro \
-v ~/.devtrace:/root/.devtrace \
-e DEVTRACE_GITHUB_AUTHOR=your-username \
-e DEVTRACE_GITHUB_OWNER=your-org \
ghcr.io/yakupkeskin/devtraceOr with Compose: docker compose up -d
- Python 3.11+
- GitHub CLI authenticated (
gh auth login) - Claude Code and/or Codex CLI in use
Copy .env.example to .env and set at minimum:
DEVTRACE_GITHUB_OWNER=your-org-or-username
DEVTRACE_GITHUB_AUTHOR=your-github-username| Variable | Default | Description |
|---|---|---|
DEVTRACE_GITHUB_OWNER |
— | GitHub org or username to scope commit search |
DEVTRACE_GITHUB_AUTHOR |
— | GitHub username for author filter |
DEVTRACE_GITHUB_SINCE_DAYS |
90 |
Days of history to fetch |
DEVTRACE_TIMEZONE |
UTC |
IANA timezone for date grouping |
DEVTRACE_IDLE_THRESHOLD_MINUTES |
30 |
Gaps longer than this excluded from active time |
DEVTRACE_MIN_SESSION_SECONDS |
60 |
Minimum session length to include |
DEVTRACE_GIT_SOLO_MINUTES |
15 |
Minutes credited per commit outside an AI session |
| Page | Description |
|---|---|
| 📊 Daily Overview | Hours, sessions, commits by day with project/source filter |
| 🕐 Timeline | Hour-by-hour activity heatmap showing when you code |
| 📁 File Activity | Most-changed files ranked by lines added + removed |
| 🪙 Token Usage | Input/output/cache tokens by model, project, and day |
| 🔍 Projects | Deep-dive into a single project's activity |
| 🎫 Tickets & Time | Jira tickets with estimated vs. logged hours |
| ⏱️ Tempo Worklogs | Weekly calendar of submitted Tempo worklogs |
| 🔗 Tempo Matching | Match sessions to tickets and bulk-submit worklogs |
devtrace today
devtrace week
devtrace week --offset -1 # last week
devtrace month --month 2025-12
devtrace sessions --date 2025-12-01
devtrace files --top 20
devtrace tokensDevtrace uses the GitHub CLI to fetch commits — no personal access token needed.
# 1. Install gh CLI (if not already)
# https://github.com/cli/cli#installation
# 2. Authenticate
gh auth loginAfter login, open the GitHub settings tab in the dashboard to select which orgs and repos to track.
- Go to Atlassian API tokens
- Click Create API token, give it a label (e.g.
devtrace) - Copy the token
- Open the Jira settings tab → enter your Site URL (
https://your-domain.atlassian.net), Email, and paste the token
- Go to Tempo API tokens (requires Tempo admin access or your team's Tempo settings page)
- Click New Token, give it a name
- Copy the token
- Open the Tempo settings tab → enter your Email and paste the token
Tokens are stored in your OS keychain (Windows Credential Manager / macOS Keychain) — never in plain text.
- Credentials are stored in the OS keychain via
keyring, falling back to the local config file only when keyring is unavailable. - Subprocess calls sanitize all user-controlled inputs — repo paths are validated against
owner/repoformat, author patterns are checked for flag injection. - JQL queries validate project keys with a strict
^[A-Z][A-Z0-9_]*$regex before interpolation. - Log output never includes raw API query parameters or full stderr that could leak credentials.
uv sync --extra dev
pytest
ruff check devtrace tests