Starter page — grow this as real questions come in.
Does Nodal need access to my database?
Does Nodal need access to my database?
No. Nodal never connects to your database — in the open-source repo or the hosted MCP server.
Your agent queries your warehouse through your own read-only connector; Nodal serves the
context (definitions and canonical queries), not your data. See
what Nodal can access.
What does the hosted MCP server need, then?
What does the hosted MCP server need, then?
GitHub access, nothing else. Your analytics-context repo gets read + write — read to
serve your context, write so your analyst can edit it and open pull requests from the Nodal
admin (every change lands as a reviewable PR). Your dbt project repo, if you connect it for
lineage, stays read-only — Nodal can never edit it. You grant this by installing the
Nodal GitHub App (the default) or by pasting a fine-grained token — see
Connect GitHub.
Is the context format proprietary?
Is the context format proprietary?
No. The Analytics Context Format (ACF) is open (Apache-2.0), and the eval harness also reads
dbt models/docs and plain markdown. You can self-host and never talk to us.
How do I run Nodal in my own cloud/VPC?
How do I run Nodal in my own cloud/VPC?
For data-residency or security requirements, contact sales at
info@nodaldata.io.