> ## Documentation Index > Fetch the complete documentation index at: https://docs.nexalytica.com/llms.txt > Use this file to discover all available pages before exploring further. # Nexalytica Chats: Conversations, Sources, and Responses > Find, create, and manage chats in Nexalytica. Read AI responses with citations, data chips, and tool timelines, and act on messages. Chats are your primary way to ask questions about your data in plain language and receive AI-generated answers, charts, and tables. Each chat holds a full conversation thread with one AI agent, and you can attach data sources or files at any time to focus the agent's responses. This page walks you through managing your chat list, working inside a conversation, and understanding everything the assistant surfaces in its replies. ## Finding and managing chats The chat list is the starting point for all your conversations. It displays every chat you own or that has been shared with you. ### Chat list controls | Control | What it does | | ----------------------- | ----------------------------------------------------------------------- | | **New Chat** button | Creates a fresh conversation and opens it immediately | | **Search box** | Filters the list in real time as you type | | **Row click** | Opens the chat in the current tab | | **Row middle-click** | Opens the chat in a background tab | | **Row ⋮ menu → Edit** | Renames the chat | | **Row ⋮ menu → Share** | Opens the share dialog | | **Row ⋮ menu → Delete** | Moves the chat to Trash (soft-delete; permanently removed after 3 days) | | **Create Chat** button | Appears on the empty playground; same as New Chat | Deleted chats stay in Trash for **3 days** before being permanently removed. During that window you can restore them from the Trash view. ### Creating a chat Click **New Chat** in the sidebar or **Create Chat** on the empty playground. A blank conversation opens immediately with the default agent pre-selected. Type your first question in the composer at the bottom and press **Enter** to send. *** ## Inside a conversation Once you open a chat, the header gives you quick access to settings that affect the whole thread. The composer at the bottom is covered in detail on the [Composer & Tools](/chats/composer-tools) page. ### Header controls | Control | What it does | | -------------------- | --------------------------------------------------------------------------------------------------------------------------------- | | **Chat title** | Click to edit inline; press **Enter** to save or **Escape** to cancel. A pencil icon appears on hover to signal it is editable. | | **Agent name label** | Shows which AI agent is currently answering in this chat | | **Sources (N)** | Opens the Sources side panel; **N** is the number of active sources | | **Artifacts** | Opens the Artifacts side panel; displays a count badge when files are attached | | **Export Chat** | Generates and downloads a PDF of the full conversation (title, agent, and message history); shows **Exporting…** while processing | | **Share** | Opens the share dialog | ### Sources panel The Sources panel lets you choose which connected data sources the agent can read in this conversation. * Click **Sources (N)** in the header to open the panel. * Tick a source to attach it; untick to detach it. * The count **N** in the button updates to reflect the number of currently active sources. Attaching only the sources relevant to your question keeps responses faster and more focused. ### Artifacts panel The Artifacts panel manages files you upload directly into the chat for one-off analysis. Click the **Artifacts** button in the conversation header. Drag and drop a file onto the upload area, or click it to open the file picker. Supported formats: **PDF, Excel, CSV, and plain text**, up to **100 MB** each. Each file in the panel exposes the following actions: | Action | What it does | | ------------- | ------------------------------------------------------------------------------------- | | **Preview** | Opens an in-app preview of the file | | **Promote** | Saves the file as a reusable Data Source or File resource available across Nexalytica | | **Download** | Downloads the original file to your device | | **Delete** | Removes the file from this chat | | **Load more** | Loads additional attachments when the list is long | Click **Close** or click the **Artifacts** button again to dismiss the panel. *** ## Reading assistant responses The assistant's replies contain several interactive elements designed to help you verify data and explore results. ### Citations When the agent draws on a specific source to support a claim, it inserts a numbered marker such as **\[1]** or **\[2]** inline in the text. Click any marker to open a popover that identifies the source behind that claim. ### Data-value chips Key figures — dollar amounts, percentages, and other important metrics — are highlighted as tinted chips in the response text. These chips make it easy to scan a reply and spot the numbers that matter most. ### Tool and step timeline When the agent calls external tools or works through multiple reasoning steps, those steps appear as a collapsible group in the reply. * **While the response is generating**, the timeline expands so you can watch the agent's progress in real time. * **Once the response is complete**, the group collapses to a single-line summary. Click it to expand the full list of steps again. ### Inline charts, tables, and dashboards The agent can generate visualisations — charts, tables, and mini-dashboards — directly inside the conversation. These visuals are interactive: hover for tooltips and click data points where supported. ### Turn timer Each assistant reply shows how long the agent took to generate the response. Use this to gauge the cost of complex queries and to decide whether a different [Deep Thinking level](/chats/composer-tools#deep-thinking-modes) would be more appropriate. *** ## Message-level actions Every assistant message includes two action buttons below it. | Action | What it does | | -------------- | ---------------------------------------------------- | | **Copy** | Copies the full assistant reply to your clipboard | | **Regenerate** | Re-runs the same request to produce a fresh response | Regenerating does not delete the previous response from the thread; it appends a new reply so you can compare outputs. *** ## Permissions Different roles have different levels of access within a chat. * Read the conversation thread * Copy assistant replies to clipboard * Regenerate any assistant message * Export the conversation as a PDF * View the Sources and Artifacts panels Everything above, plus: * Rename the chat (via the row ⋮ menu or the inline title) * Attach and detach data sources * Upload, manage, and promote artifacts * Re-share the chat within their permission level Everything Editors can do, plus: * Delete the chat to Trash Admins see only their own chats and chats shared with them in the standard list — there is no separate admin view for the chat list.