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

# Search conversations

> Find an agent's existing conversations by words in a question or an answer. Search helps when you remember the topic but not the date or title.

Find an agent's existing conversations by words in a question or an answer. Search helps when you remember the topic but not the date or conversation title.

Search covers the selected agent's accessible conversation messages. It does not search other agents or the documents in your Library.

**Already on the right agent's Analytics page?** Continue with [Search the messages](#search-the-messages).

## Open the conversation list

1. Open [Analytics](https://app.alysium.ai/analytics).
2. Under **Your Agents**, select the intended agent.
3. Open **Conversations**, or **Chats** on a smaller screen, if the list is hidden.
4. Continue with [Search the messages](#search-the-messages).

The link opens the Analytics overview, where you choose your own agent. If it asks you to sign in, sign in. Open Analytics again if needed.

## Search the messages

Start with one distinctive word likely to appear in the exchange. More words can narrow the search, so begin simply if you are unsure of the wording.

1. Under **Recent Conversations**, select **Search conversations...**.
2. Type your search word.
3. Wait for the searching indicator to finish.
4. Select a matching conversation.
5. On a small screen, open **Msgs** if needed.
6. Read the messages to confirm that you found the exchange you wanted.

Typing triggers the search. There is no separate Search button, and you do not need to send a chat message.

For example, suppose a visitor to a fictional outdoor-shop assistant asked what to bring for a return. The reply mentioned a receipt. Searching for **receipt** can find that conversation even though the visitor did not type the word, because search includes the agent's replies.

## Understand which conversations are searched

**Searching all conversations** means the search covers all time, rather than the summary's selected period or a custom list date range. A **Last 7 days** summary does not restrict keyword results to that week.

Search still follows the list's **Show teammate conversations** setting, where available. Turn it on when looking for your own practice chats. This setting includes you as the space owner; your space currently has no other members. It changes the list scope without changing visitor summary metrics.

## Return to the ordinary or dated list

1. Select the X in the search field, labelled **Clear search**.
2. Wait for the regular conversation list to return.

To choose dates, continue with [Filter conversations by date](https://docs.alysium.ai/analytics/filter-conversations-by-date.md). Search can hide the date picker while it is active.

Searching and clearing search change the view without editing or deleting messages. No Save action is needed. Selecting a result can mark that conversation as read.

## If nothing matches

1. Check the selected agent's name.
2. Check your spelling.
3. Try one distinctive word instead of a complete sentence.
4. Turn on **Show teammate conversations**, if available, when looking for your own practice chat.

A topic in a knowledge document will not appear here unless it was also mentioned in an accessible conversation message. Try a different word from the exchange if your first search returns no matches.

## If searching fails

1. Select **Retry** if the failed-read message offers it.
2. [Clear the search](#return-to-the-ordinary-or-dated-list) to check whether the regular list loads.

If the problem continues, record the agent name, search word and displayed error for support. Avoid including private transcript contents unless they are needed to explain the problem.

Once you find the conversation, use [Export conversations](https://docs.alysium.ai/analytics/export-conversations.md) for a local copy or [Read answer feedback](https://docs.alysium.ai/analytics/read-answer-feedback.md) to investigate a reported answer problem when Feedback is available.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.