Claude Code Karma CLI
A local dashboard for Claude Code sessions: timelines, token costs, tools, agents, skills and live sessions
- 337 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Reads full session transcripts that may contain secrets
- Installs Claude Code hooks and can add a login item
Install
- Works in
Manual install
git clone https://github.com/JayantDevkar/claude-code-karma.git
cd claude-code-karma/api
pip install -e ".[dev]" && pip install -r requirements.txt
uvicorn main:app --reload --port 8020In a second terminal: cd frontend, npm install && npm run dev, then open http://localhost:5180.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Reads full session transcripts that may contain secrets
- Installs Claude Code hooks and can add a login item
This is third-party code. Review the repository files before installing.
What it does
Karma reads Claude Code session files from ~/.claude and shows them in a web UI on your machine. Each session shows a timeline of prompts and tool calls, tasks, file operations, subagents, skills, background shells and token costs. Global analytics cover cost, cache, models and working rhythm per project, with separate views for tools and MCP, agents, hooks and plugins. A live tracking hook shows active sessions and jumps to their terminal, and the link-ticket-to-session skill ties a session to a Linear, Jira or GitHub ticket.
Who it is for. For heavy Claude Code users who want a clearer view of costs and session history than raw JSONL.
Good fit when
- You want to see where tokens and money go
- You need to find an old session and see what the agent changed
- You want to watch several live sessions at once
Not a fit when
- You do not use Claude Code
- You need server-side team analytics: the dashboard is local
Example request
/link-ticket-to-session ABC-123Limitations
Requires Python 3.10, Node.js 18 and running two servers or the app launcher. Claude Code keeps sessions for about 30 days, and older ones disappear from the dashboard too. Live tracking needs hooks installed per SETUP.md. Costs are estimated from local data and may differ from your bill.
How to disable. Stop the servers, remove Karma hooks from Claude Code settings and the launcher with python3 scripts/install_karma_app.py --uninstall.
FAQ
Does Karma send data to the cloud?
No, per the README it runs locally with no accounts or telemetry and only reads files from ~/.claude.
Does Karma change Jira or Linear tickets?
No, it stores only the link and cached title and status, and never writes to the tracker.
Discussion
No questions or reviews yet.
Related
CLI toolsA local real-time dashboard for Claude Code and Codex sessions: agent activity and tool calls
Local search across your whole Claude Code session history: decisions, commands, variable values and code you already wrote, through 13 MCP tools
An environment for Claude Code that keeps context across sessions via hooks, ledgers and handoffs, orchestrates agents and saves tokens
A CLI and the aitmpl.com catalog: installs agents, commands, MCP servers, hooks, settings and skills into Claude Code, plus session analytics
A self-improvement plugin for Claude Code, Codex and OpenCode: turns corrections into rules the agent follows in later sessions
A learning kit for Claude Code: feature walkthroughs plus copy-paste templates for commands, skills, memory, hooks and subagents
README in short
The README walks through dashboard sections with screenshots: sessions, session tabs, projects, analytics, tools, agents, hooks, plugins, skills and ticket linking. Quick start runs a FastAPI backend and a SvelteKit frontend, pointing to SETUP.md for live tracking. A launcher installer for macOS and Windows supports autostart. Apache 2.0 licensed.