Skip to main content
Use Case · LlamaIndex

Your LlamaIndex pipeline connects to everything. Civic decides what it can see.

LlamaIndex makes it easy to build RAG pipelines over any data source. Civic makes sure the pipeline only indexes what you intended.

use case / llamaindex

The RAG pipeline indexed the finance folder. Quarterly earnings leaked to the chatbot.

Just imagine, one day…

You build a LlamaIndex pipeline to power an internal Q&A chatbot. It connects to Google Drive and indexes folders for engineering, product, and marketing. The connector has read access to the entire drive.

A shared parent folder also contains the finance team's directory. The pipeline indexes it without distinction. Someone asks the chatbot about the product roadmap and gets an answer that cites unreleased quarterly earnings data. Investor relations finds out from a screenshot in Slack.

A RAG pipeline that indexes everything doesn't have a search problem. It has a permissions problem.

Without folder-level scope, your pipeline indexes everything the credentials allow.

It does things you did not intend

You wanted engineering and product docs. The pipeline indexed finance too, including unreleased earnings and confidential board materials.

You cannot prove what happened

Which folders were indexed? Which documents were cited in chatbot responses? Without audit logs, you're grepping vector store metadata.

You cannot stop it fast enough

The data is embedded in the vector store. Deleting the pipeline doesn't delete the embeddings. Confidential data is still searchable.

It gets confused and you never know

A file titled “Product Strategy Q3.pdf” in the finance folder. The pipeline indexed it as product content. It was the board's acquisition strategy.

Connect LlamaIndex through Civic in three steps

LlamaIndex Agent

Add Google Drive.

Scope to /shared/engineering and /shared/product.

Make /shared/finance invisible.

Done. 1 tool connected:

✓ Google Drive — engineering + product only

Your MCP URL and token are ready to copy.

The same scenario. Different outcomes.

Without Civic, the pipeline indexes every folder. With Civic, it only sees the folders you scoped.

$ python rag_pipeline.py --index
[llamaindex] connecting to Google Drive...
[drive:read] indexing /shared/engineering — 520 files
[drive:read] indexing /shared/finance — 89 files
[drive:read] ✘ Q3_earnings_draft.xlsx indexed
$ # finance data in the vector store. anyone can query.
$ python rag_pipeline.py --index
[llamaindex] connecting to Drive via Civic...
[drive:read] indexing /shared/engineering — 520 files
[drive:read] indexing /shared/product — 310 files
[nexus] /shared/finance — not in scope
$ # only engineering + product indexed. finance invisible.

Ship safer with Civic

We'll help you implement authenticated, scoped, and auditable access without slowing down your build.