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.

You can attach files to any chat message and ask the agent to analyze, summarize, transcribe, or reason about them. Files are uploaded securely before your message is sent, then passed to the AI model alongside your text.
Vision and multimodal analysis depend on the model configured for the agent. Models like GPT-4o, Claude 3, and Gemini natively understand images and can describe, compare, or read text from them. If the agent is running a text-only model, file analysis may be limited to text extraction from documents.

​
Attaching a file

1

Open the file upload panel

Click the paperclip icon to the left of the message input. The upload panel expands above the input box. You can also drag and drop a file directly onto the chat window at any time.
2

Select or drop your file

Click Attach Files to open the file picker, or drag your file into the dashed drop zone. You can attach up to 5 files per message (the limit may vary depending on your deployment).
3

Wait for the upload to finish

A progress bar tracks each file. The status changes from Uploading → Processing → Ready to send. Do not send the message until all files show Ready to send.
4

Type your message and send

Add any instructions in the message box — for example, “summarize this document” or “what does the chart in this image show?” — then press Enter to send. You can also send a file with no message text if you just want the agent to respond to what it sees.
A preview of each attached file appears in the conversation thread alongside your message.

​
Removing a file before sending

To remove an attached file before sending, click the × button next to the file in the upload panel. The file is removed from the pending message. Already-sent files remain in the conversation thread and cannot be removed.

​
Supported file types

FormatWhat the agent can do
JPEG, JPGDescribe, compare, read text from, or analyze the image
PNGDescribe, compare, read text from, or analyze the image
GIFAnalyze frames and describe content
WebPDescribe and analyze the image
Image files show an inline preview in the conversation thread. Vision analysis requires a model that supports images (such as GPT-4o, Claude 3, or Gemini).

​
File previews in the conversation

After you send a message with attachments, each file appears as a preview card in the conversation thread:
  • Images show an inline thumbnail
  • Documents show a file icon with the filename and page count (if available)
  • Audio and video show a player or file icon with the duration
  • Data and code files show a file icon with the filename and size
Previews are shown for both your messages (as uploader) and in any follow-up reference the agent makes to the content.

​
Referencing files in follow-up messages

Files you attach are stored for the duration of the conversation. You can refer to a previously attached file in any follow-up message — for example:
“Go back to that PDF I uploaded earlier — what does section 3 say?”
The agent has access to all files attached during the current conversation and can reference them without you re-uploading.

​
File size limits

The maximum file size per attachment is set by your deployment configuration. If a file exceeds the limit, you’ll see an error message before the upload begins. Try compressing large files or splitting them into smaller parts if you hit the limit. If you attach files that are too large for the model to process, the agent will let you know in its response.