Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.brainstormer.io/llms.txt

Use this file to discover all available pages before exploring further.

The chat interface is where you talk to your agents. Each conversation has its own memory, message history, and token usage tracking, so you can have focused exchanges and always pick up where you left off.

Opening the chat interface

To start a conversation with an agent, go to Agents in the sidebar, find the agent you want to chat with, and click Chat. You can also open a specific past conversation from the Chat History page. When the chat opens, the agent sends a welcome message (if one is configured) and the input box is ready for your first message.

Sending messages

Type your message in the input box at the bottom of the chat window and press Enter to send.
ActionShortcut
Send messageEnter
Insert a newlineShift + Enter
The agent’s response streams in word by word. Once the full response arrives, the message is saved to the conversation and token usage appears below it.

Conversation memory

Within a single conversation, the agent remembers everything that was said earlier. You can refer back to earlier context — “what did you say about that?” or “go deeper on the second point” — and the agent understands what you mean. Memory is scoped to the conversation. Switching to a new conversation starts the agent with a clean context.
When you switch topics significantly, start a new conversation instead of continuing an existing one. A focused conversation gives the agent cleaner context and tends to produce better responses.

Starting a new conversation

You can start a new conversation at any time without losing your existing ones:
1

Open history or header

Inside the chat interface, click History in the top-left corner to open the conversation sidebar, or look for the New Chat button in the header if you already have an active conversation.
2

Click New Chat

Click New Chat. The agent resets to a fresh context and sends its welcome message again (if configured).
3

Type your first message

Your previous conversations are saved and accessible from the history sidebar — they’re never deleted when you start a new one.

Browsing chat history

All past conversations are saved and accessible from two places:

History sidebar

Inside the chat interface, click History to open a sidebar listing all conversations for that agent. Search, select, and continue any conversation.

Chat History page

Go to Chat History in the sidebar to see conversations across all your agents in one place, with summary stats.

Using the Chat History page

The Chat History page shows:
  • A summary of total conversations, total messages, and active conversations (last 7 days)
  • Average messages per conversation across your agents
  • A searchable list of all conversations for the selected agent
If you have more than one agent, use the agent selector buttons at the top of the page to filter conversations by agent. Click any conversation in the list to read through its full message history. From that view, click Continue Chat to resume the conversation in the chat interface.

Searching and managing conversations

In the conversation list — whether in the sidebar or Chat History — you can:
  • Search conversations by typing in the search box. Results filter as you type.
  • Rename a conversation by clicking the pencil icon. Press Enter or click away to save the new title.
  • Delete a conversation by clicking the trash icon. You’ll be asked to confirm.

Multiple concurrent conversations

You can have multiple open conversations with the same agent simultaneously. Each conversation maintains its own independent memory and message history. This is useful for exploring different use cases or comparing different lines of reasoning without mixing contexts.

Token usage

After each agent response, a usage bar appears below the messages showing:
  • The number of tokens used by the last response (input + output combined)
  • The credit cost for that response
This helps you understand how much each interaction costs and spot unexpectedly large responses.

Citations from knowledge bases

If the agent has a knowledge base linked, its responses may include numbered citation markers like [1]. Hover over a citation to see:
  • The name and source of the knowledge base document
  • A relevance score showing how closely the source matched your question
  • A preview of the source text
  • A link to view the original source
Citations only appear when the agent retrieves content from its knowledge base to answer your question. See the Knowledge Base documentation to learn how to attach a knowledge base to an agent.