Snowflake and Tray

Snowflake is the data cloud where teams store, transform, and query information across databases, schemas, and warehouses. But on its own, it isn’t built to coordinate actions across external systems or trigger workflows beyond data operations. Tray fills that gap. With Tray, query results, Snowpipe loads, and task runs can initiate automations and agents that update analytics tables, refresh pipelines, or sync downstream apps, writing outcomes back to Snowflake with accuracy and control.

With Tray, you build
Automations
Integrations
 and
Agents
 that connect Snowflake across your enterprise systems and workflows.

Automations

Automations with Snowflake and Tray

Tray turns Snowflake activity into action. Query results, stream deltas, Snowpipe loads, and scheduled checks trigger flows that apply rules and approvals, then write outcomes back to Snowflake or connected systems. Status updates post to the right channels or records so owners can act, and every step stays traceable and auditable.

Integrations

Integrations with Snowflake and Tray

Tray connects Snowflake to the rest of your enterprise stack so data, queries, and events move with purpose. These integrations power the automations below and the tools your agents use, turning data flows into business actions across analytics, operations, and finance.

Snowflake integration capabilities

Integrate Snowflake with 700+ applications, plus any system with an API using Tray’s HTTP connector. These domains mirror Snowflake’s API and how teams actually work in Snowflake.

Turn SQL into action and make results visible everywhere

  • Submit & poll statements: Run queries via the SQL API, fetch status and results, including partitioned result sets
  • Parameterized runs: Pass safe inputs for repeatable transformations and checks
  • Long-running jobs: Execute asynchronously, retrieve results and post summaries where teams work
  • Result routing: Write outputs to tables, files, or downstream apps

Agents

Build agents with Snowflake and Tray

Tray agents bring Snowflake data into the flow of work. Agents query scoped datasets, trigger governed actions, and write outcomes back to Snowflake tables and connected apps, all from chat or the web.

Ground answers in approved data, not guesswork

  • Scope: Index selected schemas, views, and docs you authorize, or stay stateless and query live
  • Freshness: Re-ingest on a schedule, set retention windows
  • Linked knowledge: Optionally include data catalog notes and runbooks

Tray vs Snowflake Cortex Agents

Cortex Agents are Snowflake’s native agent option. If most of your processes and data live in Snowflake, Cortex Agents may fit. If you need a Snowflake agent that operates across the rest of your stack, Tray is the better choice.

Merlin Agent Builder

Acts across Snowflake and 700+ external systems via connectors/HTTP and writes back to Snowflake and third‑party apps

Scopes to specific schemas/tables/views, refreshes by scheduled queries or event triggers, and stores run context and results for reuse

Supports your own LLMs or Tray-hosted models for flexible deployment

Enforces cross-system guardrails, RBAC, approvals, and end-to-end execution logs

Deploys in Slack/Teams/web/HTTP, and orchestrates multi-app workflows triggered by messages, schedulers, or webhooks

Snowflake Cortex Agents

Operates within a Snowflake account and primarily queries/manipulates Snowflake data and registered tools

Stays grounded in Snowflake objects and functions registered to the agent while data remains inside Snowflake

Uses models hosted/managed by Snowflake for agent reasoning

Enforces Snowflake roles/policies with governance focused inside Snowflake

Runs within Snowflake context surfaced through Snowflake interfaces or downstream apps wired to Snowflake

Requires Snowflake account and configuration

FAQs

Yes. Combine Snowflake streams (for change data) and tasks (for scheduling) with Tray triggers to run downstream workflows on data arrival. You can also check Snowpipe load history to confirm freshness before taking action.

What comes standard with Tray

Whether your systems, data, or models run in the cloud or on-premises, Tray connects them in one secure platform. Every connection, workflow, and agent operates under IT governance with encryption, audit logging, and access controls built in. Security teams can trust that all integrations comply with enterprise network and authentication policies.

Universal connectivity

  • Prebuilt connectors: 700+ connectors plus a universal HTTP connector for any REST API
  • Custom connectors: Build custom connectors that behave like native ones
  • Connect anywhere: Cloud or on-prem systems supported

Learn more about our connectivity options

On-premises connectivity

  • Connect securely: Access on-premises systems, whether first-party or third-party
  • Meet network requirements: Connect through approved configurations that align with enterprise security policies
  • Enterprise protocols: Support multiple on-premises security standards for safe integration

Learn more about on-premises connectivity

Authentication management

  • Secure credentials: Collect and store authentications with full encryption
  • Encrypted data: Protect all data at rest and in transit
  • Role-based control: Partition credentials by workspace and access level

Learn more about authentication management

Security and governance

  • Certified compliance: SOC 2 Type II, GDPR, CCPA, HIPAA
  • End-to-end protection: Encryption, detailed audit logs, scoped connections, and OAuth scopes

Learn more about security and governance