viberank MCP
An MCP server and CLI that submit ccusage stats to a shared AI coding spend leaderboard and let the agent read the board.
- 121 ★ on the repository
- Last commit
Extended access
Why this level
- Spend stats go to an external site and appear in a profile under your GitHub name.
- The server runs `ccusage` through npx on your machine.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/viberankDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Spend stats go to an external site and appear in a profile under your GitHub name.
- The server runs `ccusage` through npx on your machine.
This is third-party code. Review the repository files before installing.
What it does
viberank is a leaderboard of AI coding spend and tokens across Claude Code, Codex, Gemini CLI, Copilot, OpenCode and other tools that ccusage reads. The `viberank-mcp` package connects it to an agent with three tools. `get_usage` runs `ccusage@20 daily --json --by-agent` locally and returns a summary without sending anything. `submit_to_viberank` uploads data to the leaderboard and returns a profile link, rank and README badge markup. `get_leaderboard` reads the board with a per-tool filter, up to 50 rows. The same board accepts data through `npx viberank-cli`, a `curl` request or a web upload. `npx viberank-cli autosubmit` registers a daily submission with the system scheduler.
Who it is for. Claude Code and Codex developers who want to compare their agent spend with others and keep a usage history.
Good fit when
- Claude Code deletes old transcripts after 30 days and the spend history must survive.
- You are curious how your Codex spend compares with other users.
- Work happens on two computers and the stats should sum into one profile.
- A README badge with the leaderboard rank is wanted.
Not a fit when
- Model spend must stay private: a leaderboard profile is public.
- You only need local analytics without sending data to an external site, where ccusage alone fits.
- A GitHub account must not be tied to usage data.
Example request
Show my ccusage spend summary without submitting to the leaderboard, then list the top 10 for Claude Code.MCP
- Transport
- stdio
- Authentication
- key is optional
Environment variables
VIBERANK_TOKENsecret- viberank API token (starts with `vbr_`): submissions go under your account and are marked verified.
Limitations
Submitting creates a public profile under a GitHub username; without a token the agent first asks you to confirm your handle. The server runs `ccusage@20` through npx, so Node.js 18 or newer and npm access are required. Alongside totals, the count and size of transcript files per month and a hash of the folder path are sent, but not their content. The board rejects suspicious figures under the rules in `VALIDATION.md`, with total cost capped at 5,000 dollars for each of 365 days.
How to disable. Remove the entry with `claude mcp remove viberank`. `npx viberank-cli autosubmit off` stops the scheduled submission, and `npx viberank-cli logout` also removes the token. You can revoke the token in the site settings.
FAQ
How do I install viberank in Claude Code?
Run `claude mcp add --scope user viberank -- npx -y viberank-mcp`. On native Windows put `cmd /c` before `npx`.
Is a token required?
No. With a `VIBERANK_TOKEN` token the submission goes under your account and is marked verified; without it the agent confirms a GitHub handle.
Is conversation content submitted?
No. Daily and per-model totals are sent, and for transcripts only the file count, size and a hash of the folder path.
How do I submit every day?
Run `npx viberank-cli login`, then `npx viberank-cli autosubmit`. The job is registered with launchd, systemd or Windows Task Scheduler.
Which tools does the server have?
`get_usage` shows a local summary, `submit_to_viberank` uploads data to the board and `get_leaderboard` reads the table, up to 50 rows with a per-tool filter.
Discussion
No questions or reviews yet.
Related
MCP serversCLI that counts token usage of Claude Code, Codex, Cursor and other agents from local logs and syncs it to vibecafe.ai.
Token and cost reports for Claude Code, Codex, Gemini CLI and other agents from local logs, by day, month and session
A CLI and MCP server that tally token spend across 41 tools and agents, break it down by project and model, and help cut wasted tokens
Splitrail tracks tokens and cost for Claude Code, Codex, Cline, OpenCode and more agents from local logs and serves them over MCP
A Windows and Linux tray app shows Claude rate limits and reports the remaining quota to an agent via a built-in MCP server.
A limits and token spend monitor for Claude Code, Codex, Antigravity and Grok CLI: macOS menu bar, Windows tray, HTML reports
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| viberank (this card) | no | key is optional | 121 | |
| ccusage | no | 19,337 | counts spend locally; viberank takes its data and publishes it to a shared leaderboard | |
| Claude Usage | no | 2,286 | a local dashboard over Claude Code logs; viberank compares users with each other |
README in short
The leaderboard publishes profiles, README badges, tool comparisons and a subscription calculator. Costs are computed at API list price, so on a subscription they are a notional estimate.