Box files
Feed type: Site / Box
Folders and nested files
PDFs, Office files, and uploads
Shared folders across the enterprise
File metadata and version history
Setup fields
FolderAllowed PathsExcluded PathsMaximum Files
Ingest
What gets ingested
Graphlit reads the selected Box surfaces, converts the source material into searchable content, and keeps enough source metadata for grounded answers, retrieval, and downstream agent workflows.
Box
Prepare
What Graphlit prepares
The integration is not just a connector. Graphlit turns Box into content that can be searched, cited, summarized, enriched with observations, and supplied to agents through the Graphlit API and MCP.
{
"source": "Box",
"content": "Box files",
"extracted": [
"entities",
"observations",
"facts",
"summaries"
],
"citations": [
"Box source references"
],
"agentContext": {
"searchable": true,
"groundedAnswers": true,
"deliveryEnabled": false
}
}Use cases
What you can build with Box
01
Build a Box knowledge base from folders and nested files and pDFs, Office files, and uploads and search it alongside the rest of your Graphlit content.
02
Ask grounded questions over Box context without forcing users to leave the tools where the work originally happened.
03
Extract people, organizations, projects, events, facts, and decisions from Box content so agents can reason over durable context.
04
Combine Box with Slack, Notion, and OneDrive to build cross-tool RAG, research, and operational intelligence workflows.
Need a code sample? Ask our AI assistant
SDK:
Try:or
FAQ
Common questions
Related
Commonly paired integrations
Start building
Start building with Box and Graphlit
Create a free Graphlit project, connect Box, and turn real operational context into retrieval, knowledge graphs, and MCP-native agent workflows.