{% extends 'base.html' %} {% block title %}Usage & memory analytics - Memorizz{% endblock %} {% block head %}{% endblock %} {% block content %} {% from '_ui.html' import page_header, observability_tabs %} {% from '_fmt.html' import tape %} {% call page_header('Usage & memory analytics', 'See where recorded tokens, charges, prompt-cache reuse and retrieval time go.', 'Operations', nav=observability_tabs('usage', {'traces': trace_url})) %}Export JSON{% endcall %} {{ tape(usage_tape, 'Usage summary for the filtered window', 'usage-kpis') }}
{% if preserved_filters %}Trace selection retained: {% for key, value in preserved_filters.items() %}{{ key }}={{ value }} {% endfor %}. Clear trace selection
{% endif %} {% with usage_tape_shown = true %}{% include '_usage_panel.html' %}{% endwith %} {% if coding_agents %}Codex and Claude Code sessions the MemoRizz plugin recorded, read from each agent's own session log and counted apart from the traced usage above. Cost is Claude Code's own total, or OpenAI list rates for Codex when its model has one; a ChatGPT plan isn't billed per token. Open them on Agent harnesses.
| Agent | Model | Sessions | Turns | Tokens in | Cached | Tokens out | Spend |
|---|---|---|---|---|---|---|---|
| {{ row.label }} | {{ row.model or '—' }} | {{ '{:,}'.format(row.sessions) }} | {{ '{:,}'.format(row.turns) }} | {{ '{:,}'.format(row.input_tokens) }} | {{ '{:,}'.format(row.cached_tokens) }} | {{ '{:,}'.format(row.output_tokens) }} | {% if row.priced %}${{ '{:,.4f}'.format(row.cost_usd) }}{% if row.priced < row.sessions %} + {{ row.sessions - row.priced }} unpriced{% endif %}{% else %}Unpriced{% endif %} |