Ollama Usage
Shows what is left in the Ollama Cloud monthly credit pool, with its reset date and the money already spent, in bb's usage panel. The panel reads only the OpenCode Go and Claude Code subscriptions, so an Ollama Cloud plan has no row there; this adds the pool as a dollar window beside them. Spend for the last day, week or month comes with a per-model breakdown, and bb ollama-usage prints the same numbers in a terminal. Needs an Ollama Cloud account with an API key on the host, either OLLAMA_API_KEY or the sign-in OpenCode already stored.
Overview
View sourceWhat you get
BB's Provider usage view knows about the OpenCode Go subscription, Claude Code and Kimi Code. An Ollama Cloud plan has no source there, so the panel shows nothing about it even though every agent run spends from that pool.
This plugin adds the missing source. The Ollama Cloud row shows the monthly credit pool as a dollar window: how much of the allowance the billing period has consumed, the money left, and the date the pool resets. A plan still on the old session and weekly limits appears as the same windows that plan actually uses.
What it reads
Ollama's own API answers two questions and the plugin asks both: /api/balance
for the included pool and the purchased credits, and /api/usage for spend and
request counts over the last 24 hours, week or month, with a per-model
breakdown.
Command
bb ollama-usage show [--host <id>] [--range 24h|7d|30d] [--json] prints the
same numbers in a terminal: the percentage and money of the pool, the reset
date, purchased credits, spend for the range and the models that caused it.
--json emits the full reading for a script.
Requirements
An Ollama Cloud account and an API key on the host: either OLLAMA_API_KEY, or
the Ollama sign-in that OpenCode already stores. The key is never logged or
shown, and the only host it is sent to is ollama.com. bb 0.49 or newer, and the
opencode ACP provider enabled: with no such provider no model spends credits, so
the row is not advertised at all.
What it does not do
The plugin does not show concurrency limits, because the API does not expose the number of in-flight requests. It does not manage the plan, buy credits or switch models — it reports what Ollama reports and nothing more.
More from Vladislav Chmelyuk
More in Token Usage & Limits 41
Understand or control token, context-window, and provider-quota use.



