PostgreSQL

Unified database plugin for PostgreSQL — run raw SQL or structured CRUD operations (query, find, insert, update, delete) from a single node.

Use PostgreSQL when you need to:

• Read or write data in a PostgreSQL database from a workflow
• Run parameterized raw SQL for complex queries or joins
• Perform structured CRUD without writing SQL (find, insert, update, delete)
• Give an AI Agent controlled database access via the Tools handle

The result (accessed as {{ postgres.* }}) shape:

{
  "rows": [{ "id": 1, "email": "[email protected]" }],
  "affectedRows": 0,
  "success": true
}

Add your connection string on the Env Vars page:

The plugin uses asyncpg to connect and execute queries.

Find a user by email after a webhook trigger:

1. Add a PostgreSQL node after your trigger.
2. Set Operation to find.
3. Set Table to users.
4. Set Where to { "email": "{{ webhook.body.email }}" }.
5. Set Columns to id, email, name.
6. Set Limit to 1.

Access results downstream:

{{ postgres.rows[0].id }}
{{ postgres.rows[0].email }}
{{ postgres.success }}

Insert a row from webhook payload:

• Operation: insert
• Table: orders
• Data:

{
  "customer_id": "{{ webhook.body.customerId }}",
  "amount": "{{ webhook.body.amount }}",
  "status": "pending"
}

Connect the On Error branch to handle connection failures or constraint violations.

import webhook from @tensorify/webhook-trigger:4.0.0
import db from @tensorify/postgres:2.0.0

node trigger @tensorify/webhook-trigger:4.0.0 {
    path = "/signup"
    method = "POST"
}

node find_user @tensorify/postgres:2.0.0 {
    operation = "find"
    table = "users"
    where = "{ \"email\": \"{{ webhook.body.email }}\" }"
    columns = "id, email, name"
    limit = 1
}

trigger.payload -> find_user.input

PostgreSQL has an On Error control output branch. Enable it via the "Error output handle" toggle at the bottom of the settings panel.

The error branch fires when:

• The database connection fails or DATABASE_URL is missing
• SQL syntax errors or constraint violations occur
• update or delete is attempted without a non-empty where clause
• An operation is rejected because it is not in Allowed Operations (Agent) when used as an agent tool

Connect the error branch to a fallback notification or logging action to handle failures without halting the entire workflow.

• Update and delete require WHERE: Updates and deletes without conditions are rejected at runtime. Always provide a non-empty where object.
• Positional parameters for raw SQL: Use $1, $2, … in sql and pass values as a JSON array in params — not string interpolation for user input.
• Missing DATABASE_URL: The node fails at runtime if the env var is not set. Add it on the Env Vars page.
• Agent permissions: Restrict Allowed Operations (Agent) for read-only agent access (e.g. allow only find and query).
• Check postgres.success: On the main output path, verify {{ postgres.success }} in a downstream If node if you need to branch on failure without using the error handle.

• AI Agent — connect PostgreSQL as a database tool for agents
• Webhook Trigger — trigger database writes from incoming events
• Transform — shape trigger data before insert or update
• If — branch on query results or success status
• Environment Variables — store your DATABASE_URL