Quickstart guide

In this guide, you will configure a Prequel export from source to destination. Initial setup typically takes 1-2 hours to complete.

Select your onboarding experience

Get started with the Prequel agent skill, the Prequel MCP server, a Postman collection, or curl commands.

Agent skill
MCP server
Postman
curl commands

Skills are an open standard for packaging reusable agent capabilities. Claude Code, Cursor, and other skill-aware agents can all load the Prequel Export skill.

Save the skill to your agent's skills directory

Open the raw skill file, copy the contents, and save the file to your agent’s skills directory. For Claude Code use ~/.claude/skills/prequel-export-quickstart/SKILL.md. For Cursor use .cursor/skills/prequel-export-quickstart/SKILL.md in your project root and reload the workspace.

Prompt your agent

Begin an agentic Prequel Export configuration using this prompt.

The Prequel docs are available as an MCP server, exposing search tools your agent can call to look up the correct API operations and request schemas as it works.

Add the Prequel MCP server to your tool

Point your AI tool at https://docs.prequel.co/mcp. The setup differs by client:

claude mcp add --transport http prequel https://docs.prequel.co/mcp

For full instructions, see the MCP setup guides for Claude Code and Cursor.

Prompt your agent

Begin an agentic Prequel Export configuration using this prompt.

The MCP server is read-only documentation search and does not act on your systems. If you pair it with the agent skill, which runs real API calls, review each command before running it and confirm before pointing an automated run at a customer’s production environment.

Import the OpenAPI spec into Postman

In Postman, click Import, paste the Prequel Export OpenAPI spec URL, and confirm. Postman generates a complete collection from the spec with one request per operation.

Set your API key as a collection variable

Add a PREQUEL_API_KEY variable to the imported collection and reference it from each request’s X-API-KEY header.

Open the API reference

Browse the API reference and select the operation for the resource you want to create.

Copy the sample request body to your clipboard

Use the Copy button on each operation’s example request body, paste it into your shell as the -d payload, and substitute your values and resource IDs as you go.

Prerequisites

Log into the Admin UI

Access the Prequel Admin UI at app.prequel.co, eu-app.prequel.co, or your custom domain. You can log in using Continue with Google if your company uses Google Workspace, or a one-time passcode sent to your email. If you are unable to log in, contact the Prequel team.

Access your Prequel API key

You will need a Prequel API key. For a POC, you will use a staging key, which begins with sk_staging. The key in the X-API-KEY header routes your request to production or staging based on the key. Contact your Prequel representative if you do not have one.

Set your API key

export PREQUEL_API_KEY="<your-key>"

Confirm your source data requirements

Every source table you want to transfer needs a unique ID column (e.g. id), a last-modified timestamp column (e.g. updated_at) for change detection, and a way to filter by tenant, either an organization ID column (e.g. organization_id) or a per-tenant schema. See Change detection and Multi-tenancy.

Allowlist the Prequel static IP

If your source or destination restricts inbound traffic, allowlist the Prequel static IP before connecting. See Prequel IP addresses.

Verify Prequel API access

Confirm your key works against the API. A 200 response means your environment is ready.

Verify API access

curl -H "X-API-KEY: $PREQUEL_API_KEY" https://api.prequel.co/sources

Use the Staging environmentFor initial onboarding, use the toggle in the bottom-left corner of the Admin UI to switch to the Staging environment. Your staging API key only creates and updates resources within the staging environment.When you are ready to begin work in your production environment, contact the Prequel team to receive a Production API key and flip to the Production environment in the Admin UI.

Setup overview

A Prequel Export pipeline is built in five steps:

Connect a source so Prequel can read from your database, warehouse, or object store.
Configure your data model to declare which tables and columns are shared and how changes are detected.
Add a recipient to represent the customer you are sharing data with.
Add a destination that the recipient receives data into.
Trigger a transfer and verify data lands in the destination.

Set up your pipeline

Each call returns a resource ID in the response body. Record the IDs as you go, since later steps reference them with placeholders such as <SOURCE_ID> and <DESTINATION_ID>.

Connect a source

A source is the database, warehouse, or object store Prequel reads from. Create a new user or role in your source, granted read-only permissions on only the datasets you intend to export, to adhere to the principle of least privilege. Then use the per-vendor fields from the source configuration guides to add it. See Sources.

Admin UI
API

Open Sources → Add source, choose your vendor, enter the connection details, and run the built-in connection test before saving.

curl -X POST https://api.prequel.co/sources \
  -H "X-API-KEY: $PREQUEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "source": {
      "name": "postgres-read-replica",
      "vendor": "postgres",
      "host": "read-replica.db.example.com",
      "port": 5432,
      "database": "production",
      "user": "prequel",
      "password": "<password>"
    }
  }'

Configure your data model

A model declares which source table is shared, which columns map to the destination, how rows are filtered per tenant, and how changes are detected. Models are defined as config files: store them in a Git repository and publish them with the GitHub Action, or post them to the API directly. See Models. Posting a model validates it against the source unless skip_model_validation is set.

Is your source schema-tenanted?Prequel environments are set to multi-tenant mode by default. If your source uses one schema per customer with identical tables across schemas (schema-tenanted), contact the Prequel team to switch your environment to schema-tenanted mode.

Create model

curl -X POST https://api.prequel.co/models \
  -H "X-API-KEY: $PREQUEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": {
      "model_name": "events",
      "description": "Platform events shared with customers",
      "source_id": "<SOURCE_ID>",
      "source_table": "public.events",
      "organization_column": "organization_id",
      "columns": [
        { "name_in_source": "id", "name_in_destination": "id", "data_type": "text", "is_primary_key": true },
        { "name_in_source": "event_log", "name_in_destination": "event_log", "data_type": "text" },
        { "name_in_source": "updated_at", "name_in_destination": "updated_at", "data_type": "timestamp", "is_last_modified": true }
      ]
    }
  }'

Add a recipient

A recipient represents one of your customers. The id_in_provider_system is the tenant identifier Prequel matches against each model’s organization_column to filter rows for that customer. See Recipients.

Admin UI
API

Open Recipients → Add recipient, give them a name and their tenant identifier (id_in_provider_system), and choose which products they receive (or all of them).

Pass ["all"] for products to share every model, or a list of product names to share a subset.

Create recipient

curl -X POST https://api.prequel.co/recipients \
  -H "X-API-KEY: $PREQUEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "recipient": {
      "name": "Acme Corp",
      "id_in_provider_system": "acme",
      "products": ["all"]
    }
  }'

Add a destination

A destination is where a recipient receives their data. Use the per-vendor fields from the destination configuration guides and attach it to the recipient via the Create destination endpoint. See Destinations.

During initial configuration, we recommend first configuring an internal destination you can access, so you can review landed data as if you were one of your customers. To test a destination type you do not have access to, contact the Prequel team for a sandbox instance.

Admin UI
API

Open Destinations → Add destination, select the recipient and vendor, and enter the connection details. You can open the destination’s configuration documentation from the Documentation button in the top-right corner of the form. You can also collect destination credentials from your customer with a magic link.

Create destination

curl -X POST https://api.prequel.co/destinations \
  -H "X-API-KEY: $PREQUEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "destination": {
      "name": "acme-snowflake",
      "recipient_id": "<RECIPIENT_ID>",
      "vendor": "snowflake",
      "host": "acme.us-east-1.snowflakecomputing.com",
      "database": "PREQUEL",
      "schema": "PUBLIC",
      "user": "TRANSFER_USER"
    }
  }'

Trigger a transfer

With a source, model, recipient, and destination in place, start a transfer. The first transfer runs a full refresh; subsequent transfers sync only changed rows.

Admin UI
API

Open the destination and click Sync now to kick off a transfer.

Trigger transfer

curl -X POST https://api.prequel.co/destinations/<DESTINATION_ID>/transfer \
  -H "X-API-KEY: $PREQUEL_API_KEY"

That’s it.Query the destination to confirm data arrived. The _transfer_status table records the outcome of each model’s transfer.

Once connected, data automatically replicates to the destination on an ongoing basis, starting at the next scheduled frequency interval. To onboard each additional customer, repeat Add a recipient, Add a destination, and Trigger a transfer.

Explore integrating with your product

When you are ready to move beyond manual setup, Prequel can be embedded into your product for a self-serve, monitored integration:

React SDK

Embed destination setup and transfer monitoring directly in your product UI.

GitHub Action

Publish and validate model changes from CI instead of by hand.

Webhooks and monitoring

Get notified on transfer success and failure via webhooks, Datadog, or Slack.

Error handling

Understand error codes and surface actionable failures to your customers.

Next steps

Schema evolution

Roll out column and type changes to destinations safely.

Orchestration

Tune how often each destination transfers.

Data integrity checks

Verify source and destination row counts match.

Tenancy options

Choose how tenant data is modeled and filtered per recipient.

​Select your onboarding experience

​Prerequisites

​Setup overview

​Set up your pipeline

​Explore integrating with your product

React SDK

GitHub Action

Webhooks and monitoring

Error handling

​Next steps

Schema evolution

Orchestration

Data integrity checks

Tenancy options

Select your onboarding experience

Prerequisites

Setup overview

Set up your pipeline

Explore integrating with your product

Next steps