SnapGPT

SnapGPT uses generative AI to help you build, analyze, and describe SnapLogic pipelines in Designer, and diagnose infrastructure and execution issues in Monitor.

Models and configuration

SnapGPT uses publicly-available large language models (LLMs). To respond to prompts, it queries both LLM data and information available from SnapLogic. It anonymously sends only generic requests to the LLMs. Learn more about SnapLogic security and compliance.

SnapGPT can also use Retrieval Augmented Generation (RAG). RAG gives SnapGPT the ability to improve results based on existing assets such as pipelines, Snaps, and expressions. RAG also decreases the chance of hallucinations when an exact answer isn’t available.

An Environment admin configures SnapGPT for all users. Individual users can configure use of RAG as described in Configure SnapGPT.

Org admins can track SnapGPT adoption and engagement across the environment from the SnapGPT usage page in Monitor.

SnapGPT in Designer

From the SnapGPT panel in Designer, you can generate, analyze, describe, and refine pipelines using the following skills and features:

Generate pipeline: Generate a pipeline from natural language prompts.
Generate agent and MCP Server pipelines: Generate agent pipelines and MCP Server pipelines for AI agent integration.
Refine pipeline: Update a pipeline or Snap based on natural language prompts, including renaming Snaps or adding error handling.
Analyze pipeline: Identify potential issues and receive best-practice recommendations for your pipeline design.
Describe pipeline: Generate a natural language description of a pipeline and save it to the pipeline’s Info tab.
Generate pipeline from image: Generate a pipeline from an image of a pipeline diagram.
Get preview data insights: Analyze preview data from Snaps of a validated pipeline and surface insights into key fields, business and validation rules, and PII.
Refine Snap using SnapGPT Copilot: Get AI assistance within individual Snap configuration panels.
Generate expressions: Generate expressions in expression-enabled Snap fields and Mapper fields.
Generate SQL: Generate SQL queries in SQL-enabled Snap fields.
Ask questions: Answer questions about the SnapLogic Platform.
Select a reasoning mode: Select Standard, Think, or Plan mode or combine them to control the depth of AI reasoning for your task.

SnapGPT in Monitor

Use SnapGPT in Monitor to investigate operational issues without leaving the page you are working on. The following skills are available:

Check Snaplex health: Identify potential issues and suggested fixes on your Snaplexes from the Infrastructure page.
Analyze failed pipeline execution: Identify the root cause and suggested fixes for a failed pipeline execution from the Pipeline executions page.
Plan Snaplex capacity: Project when Snaplexes will approach capacity limits based on historical usage trends from the Insights page.

Refer to SnapGPT in Monitor for more information.

Known issues

You might experience significant latency (up to one minute) or timeout issues, depending on the complexity of the request and the capacity of the LLM provider.
As with all AI based on LLMs, SnapGPT can occasionally be unreliable or inconsistent in its responses.
Think and Plan modes lock after you send the first message in a session. To switch modes, start a new chat.
Pipeline generation and refinement in Plan mode support one active session per user at a time. Starting a new session from another browser tab shows a warning and requires confirmation before ending the existing session.