Do I need to enable the pgvector extension before connecting Gateco?

Yes. Gateco requires the vector extension to be present in the database to create vector indexes and run similarity searches. In Supabase, go to Database → Extensions, search for "vector", and toggle it on. Alternatively, run CREATE EXTENSION IF NOT EXISTS vector in the Supabase SQL editor.

Should I use the anon key or the service role key?

Use the service role key (the long secret starting with eyJ...). The anon key is restricted by Row Level Security policies, which would block Gateco from reading the data it needs to apply its own policy layer. The service role key bypasses RLS — Gateco then applies its own access policies on top.

Does Gateco replace Supabase Row Level Security?

No. Supabase Row Level Security and Gateco address different layers. RLS restricts which rows a PostgreSQL role can read at the database level. Gateco provides vector database RBAC and RAG access control at the retrieval layer — gating which search results an end user's AI agent receives based on identity policies. They are complementary: keep your existing RLS rules and add Gateco's policy enforcement on top.

Connect Supabase to Gateco

Supabase is a managed Postgres platform with built-in pgvector, auth, and realtime. Gateco connects to Supabase as a Tier 1 connector, supporting full document ingestion, retroactive registration, and RAG access control over your vector tables — adding identity-aware retrieval and vector database RBAC without modifying your existing schema.

Step 1 — Enable the pgvector extension

In the Supabase dashboard, open your project.
Navigate to Database → Extensions.
Search for "vector" and toggle it on.

Or run this in the Supabase SQL editor:

sql

CREATE EXTENSION IF NOT EXISTS vector;

Step 2 — Get your database connection string

Navigate to Project Settings → Database.
Under "Connection string", select the URI tab.
Switch to "Direct connection" (not the Supabase connection pooler).
Copy the URI. It will look like:

text

postgresql://postgres:[YOUR-PASSWORD]@db.[PROJECT-REF].supabase.co:5432/postgres

Use the direct connection string (port 5432), not the pooler (port 6543). Gateco manages its own connection pool internally.

Step 3 — Get your service role key

Navigate to Project Settings → API.
Under Project API keys, copy the service_role key (labeled "secret").
Keep this key secure — it bypasses Row Level Security.

The service role key bypasses Supabase RLS. Store it securely and do not expose it in client-side code. Gateco uses it server-side only.

Step 4 — Add the connector in Gateco

Navigate to Connectors → Add connector → Supabase.
Paste your direct connection string from Step 2.
Click Test connection.
Click Save.

Step 5 — Configure search settings

Click your Supabase connector → Search config.
Set the Vector table name (e.g. documents).
Set the Embedding column (e.g. embedding).
Optionally set Content column and additional Metadata columns.
Save.

Configuration reference

Field	Example	Description
`connection_string`	`postgresql://postgres:pass@db.xxx.supabase.co:5432/postgres`	`Direct PostgreSQL connection URI`
`table_name`	`documents`	`Table containing vector embeddings`
`embedding_column`	`embedding`	`Column with type vector(n)`
`content_column`	`content`	`Document text column (optional)`

Troubleshooting

Error	Cause	Fix
`Relation "vector_extension" does not exist`	`pgvector not enabled`	`Enable the vector extension in Database → Extensions`
`permission denied for schema`	`Using anon key instead of service role`	`Use the service_role key from Project Settings → API`
`SSL required`	`Missing sslmode parameter`	`Add ?sslmode=require to the connection string`