uni-social-tokens
v0.2.0
Published
Opt-in campus AI coding token leaderboard for uni.social
Readme
uni.social tokens
Opt in to the AI coding token leaderboard for your verified university on uni.social.
npm install -g uni-social-tokens
uni-social-tokens --install-autosyncThe first run asks for a one-time connection code from the Tokenmaxxing page. It then syncs every 15 minutes while your laptop is on, using your operating system's user-level scheduler. Changed daily totals are uploaded; unchanged totals are skipped. Later runs reuse a device credential stored at ~/.config/uni-social-tokens/device.json with owner-only file permissions. You can disconnect and delete a device's uploaded totals from the Tokenmaxxing page.
The CLI uses ccusage in offline mode to read local usage from supported AI coding tools. It uploads date, tool name, total tokens, input/output/cache tokens, estimated USD cost, and per-model token totals. Only total tokens are used for the campus leaderboard. The detailed fields power charts visible to your account only. It does not send prompts, source code, file paths, or chat transcripts. The leaderboard is self-reported local usage; it is not independently verified and should be treated as a friendly competition. Cost depends on ccusage's local pricing data and may differ from actual billing.
Preview the exact upload without connecting or sending anything:
uni-social-tokens --dry-runUse uni-social-tokens to sync immediately, --status to check the connection, and --remove-autosync to stop the scheduled job while keeping your laptop paired. Scan selected tools with --source=codex,claude. The background job runs under your own account and stops when the laptop is off; on Linux it runs while your user session is active. No uni.social server can read your laptop directly.
Inspired by whoburnedmore, which is MIT licensed. This CLI uses uni.social device pairing and a campus leaderboard. A public GitHub badge is available only after you turn it on from Tokenmaxxing.
