Connection Details
When your vector store is configured and healthy, a Connection Details panel appears below the configuration form. This section provides the information needed to connect your external RAG applications directly to your vector store.
The configuration form is available while Connection Details checks your provider. If the check fails or takes more than 10 seconds, the panel shows an error and a Retry button; you can continue editing the configuration.
What's Included
| Provider | Connection Fields |
|---|---|
| Supabase | URL, Anon Key (public), Table Name |
| PostgreSQL / pgvector | Host, Port, Database, Username, Table Name, TLS Mode |
| Pinecone | Host URL, Index Name, Namespace |
| Cloudflare Vectorize | Account ID, Index Name, description of derived Namespace |
Field Descriptions
For Supabase
| Field | Description | Security |
|---|---|---|
| Supabase URL | Your project's API endpoint | Public |
| Anon Key | Public anonymous key for client-side access | Public (safe to expose) |
| Table Name | The table storing your vectors (default: document_chunks) | Public |
The Anon Key is designed to be public and used in client-side applications. It's protected by Row-Level Security (RLS) policies. The Service Role Key (used for admin operations) is never exposed in connection details.
For PostgreSQL / pgvector
| Field | Description | Security |
|---|---|---|
| Host and Port | Public PostgreSQL endpoint | Connection metadata |
| Database and Username | Saved target database and role name | Connection metadata |
| Table Name | Table storing vector chunks | Connection metadata |
| TLS Mode | Certificate verification policy | Connection metadata |
The database password is never exposed. External applications need their own credential and must filter every query by both the displayed Organization ID and Project ID.
For Pinecone
| Field | Description | Security |
|---|---|---|
| Host | Your index's host URL | Public |
| Index | Index name | Public |
| Namespace | Project-specific namespace (e.g., project-abc123) | Public |
The Pinecone API key is never exposed in connection details. External apps need their own API key with appropriate permissions.
For Cloudflare Vectorize
Connection Details shows the account ID, index name, and a description of the organization/project namespace. It does not reveal the namespace hash or API token. A direct external client needs its own scoped Cloudflare token, the same namespace derivation as Vector Data Loader, and filters for both organization_id and project_id. The displayed fields alone are not enough to query the correct tenant namespace.
Additional Information
The Connection Details panel also shows:
- Embedding Configuration: Provider, model, and dimensions used for your vectors
- Index Stats: Vector count and health status
Integration IDs
The Connection Details panel includes an Integration IDs section that displays the unique identifiers needed when configuring external tools like n8n, Langchain, or other RAG frameworks.
| Field | Description | When to Use |
|---|---|---|
| Organization ID | UUID identifying your organization | Required for all external tool queries to scope results to your organization's data |
| Project ID | UUID identifying the current project | Required for project-scoped queries, including PostgreSQL / pgvector and Cloudflare Vectorize |
Why These IDs Matter
External tools like n8n and Langchain need to pass filter parameters when querying your vector store. Without these IDs, queries would either fail or return data from all organizations (a security issue).
Previously, users had to:
- Query the database directly
- Check browser dev tools network requests
- Inspect cookies for the project ID
Now these IDs are prominently displayed with one-click copy buttons.
Example: n8n Supabase Vector Store Configuration
When configuring the Supabase Vector Store node in n8n:
- Copy the Organization ID from Connection Details
- Copy the Project ID for the selected project
- In n8n's metadata filter, use:
{
"organization_id": "your-organization-id-here",
"project_id": "your-project-id-here"
}
Click the copy button next to each ID for one-click copying to your clipboard.
Use Case: External RAG App Integration
These details help you configure an external RAG tool (such as Flowise, LangFlow, or n8n) or a custom application. The external client must use its own provider credentials and enforce the same organization and project isolation as Vector Data Loader. Cloudflare Vectorize additionally needs the matching derived namespace.