promptmap docs

promptmap run history — generated {{ generated_at }}

{% if latest_files %}

Files — latest run {{ latest_run.short_id }}

{% for f in latest_files %}

{% if f.preview %}

{% else %}

{% endif %}

{{ f.model_name }}{% if f.label != f.name %} · {{ f.label }}{% endif %} {{ f.size }} · sha256:{{ f.sha }} {% if not f.missing %}Download{% endif %}

{% endfor %}

{% endif %}

Run History

Click a row to expand model details.

{% if runs and not tokens_reported %}

Token usage is not being reported. To track it, have llm_call in client.py return promptmap.LLMResult(text, spent_tokens=…) instead of the bare text. How to count is up to you: input + output + thinking, or the provider's total when it reports one (e.g. Gemini usage_metadata.total_token_count, OpenAI usage.total_tokens). promptmap init scaffolds a client that does this.

{% endif %}

{% for run in runs %} {% if run.results %} {% endif %} {% else %} {% endfor %}

Run ID Status Models Started Duration Tokens used From cache Est. cold start

{{ run.short_id }} {{ badge(run.status) }} {{ run.model_count }} {{ run.created }} {{ run.duration }} {{ tokens(run.tokens_used) }} {{ tokens(run.tokens_cached) }} {{ tokens(run.tokens_cold) }}

{% for r in run.results %} {# 0 ms is a real measurement (a model that made no call); — means not recorded. #} {% if r.error %} {% else %} {% endif %} {% endfor %}

Model	Type	Config	Status	Time	Tokens	From cache	Input	Output preview	Files
{{ r.model_name }}	{{ r.model_type }}	{{ r.config or '—' }}	{{ badge(r.status) }}{% if r.cached %}Cached{% endif %}	{{ (r.execution_ms ~ ' ms') if r.execution_ms is not none else '—' }}	{{ tokens(r.tokens_used) }}	{{ tokens(r.tokens_cached) }}	{{ r.input_preview or '—' }}	{{ r.error }}	{{ r.output_preview or ('—' if not r.files else '') }}	{% if r.files %}{{ file_strip(r.files) }}{% else %}—{% endif %}

No runs recorded yet. Run promptmap run first.

{% if dag_section %}

Model DAG

{% endif %}