About this skill

Solving scattered token costs

Multi-provider API costs are hard to inspect when they live in separate consoles. token-panel-ultimate stores usage from Anthropic, Gemini, OpenAI, and Manus in one SQLite database and exposes a REST API for querying and auditing spend before the bill arrives.

How it works

Multi-provider tracking: Keeps token usage records for each provider in a single local store.
Budget alerts: Sets monthly limits per provider and warns before overspending.
REST API: Listens on port 8765 and provides endpoints such as /usage, /usage/:provider, and /budget.
Transcript parsing: Extracts token counts from OpenClaw session transcripts.
Lightweight operation: Uses SQLite and can run as a systemd service without Postgres, Redis, or a cloud account.

Boundaries and caveats

It fits local servers, self-hosted operations, or small teams that need token usage monitoring, scripted queries, or cost dashboards. It is not a financial billing, invoicing, or multi-account access control system; you still need to define data sources, budget rules, and alerting behavior.

Screenshots

Use Cases

A team reconciles weekly Anthropic, OpenAI, and Gemini spend by querying usage per provider.
An operator sets monthly token budgets for self-hosted services and needs alerts before overspend.
An engineer connects a local script to port 8765 to fetch /usage and /budget for a dashboard.
During a spike review, an engineer extracts token counts from OpenClaw transcripts to verify spend.

Best For

An operations engineer responsible for multi-model API costs needs unified usage views and budget limits.
An engineer maintaining OpenClaw sessions needs to parse token counts from transcripts.
A developer building a local cost dashboard needs REST access to /usage and /budget.
An SRE managing a self-hosted LLM gateway needs a background service to record provider usage.