> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trymeridian.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompts

> Manage your prompt library and track performance by prompt and topic.

## What this is

A table view of all tracked prompts, including their **Topic**, **Visibility score**, **Share of voice**, and **Prominence**.

## Why it matters

* Prompts are the unit of “AI search demand” inside Meridian.

* You can’t measure improvement if you aren’t tracking the right prompts.

* Topics + attributes make results explainable and actionable.

## Where to find it

**Analytics → Prompt Analytics → Prompts**

## How it works

* Each prompt is run repeatedly over time.

* Meridian aggregates results into:

  * **Visibility score** (how often you’re present)

  * **Share of voice** (how much you own vs competitors)

  * **Prominence** (how high you rank when mentioned)

<img src="https://mintlify.s3.us-west-1.amazonaws.com/meridian-25fd9b88/analytics/prompt-analytics/SCREENSHOT_PROMPTS_TABLE" alt="Prompts table with filters, attributes, export, and add prompts" />

Caption: Prompts shows performance and gives you “Analyze / Edit / Delete” controls.

## How to use it

1. Set your timeframe (e.g., **Last 7 days**) and compare to **Prev. Period**.

2. Click **+ Add prompts** to add new tracked queries.

3. Use **Filters** to focus on a topic, platform (e.g., ChatGPT), or attribute set.

4. Use **Group by** to roll up by topic/attribute (if enabled).

5. Click **Analyze** on a prompt to see response-level data.

6. Use the “…” menu to **Edit** or **Delete** a prompt.

7. Use **Export** to download your prompt table.

## How to interpret results

* If Visibility is high but Prominence is low → you’re included, but low-ranked → strengthen “best answer” structure.

* If Share of voice is low while competitors are high → your competitor set is outperforming you on this query → prioritize this prompt.

* If Prominence gets worse while visibility stays flat → you’re slipping in rank → update the content/citations the model relies on.

* If one topic’s prompts are all weak → that topic lacks coverage → create a dedicated content cluster.

* If results change drastically when you toggle platform/region → treat optimization as platform-specific.

## Common questions / troubleshooting

* “Why can’t I add more prompts?” → you may be at your plan limit (e.g., 97/300).

* “Why is the prompt missing a topic?” → edit the prompt to assign a topic.

* “Why are deltas blank?” → you need a previous period with runs to compare.

## Related docs

* [Analyze prompts](/analytics/prompt-analytics/analyze)

* [Metrics glossary](/reference/metrics-glossary)

* [Citations](/analytics/prompt-analytics/citations)

* [Opportunities](/action/opportunities)