SnapGPT

SnapGPT leverages the power of generative AI to help you build, analyze, and describe SnapLogic pipelines. SnapGPT is integrated with SnapLogic Designer.

To enable SnapGPT on your production environment, contact your Customer Success Manager (CSM).

Use SnapGPT to speed and improve integration design. It can:

Generate pipelines, expressions, mappings, and SQL queries from prompts
Configure a Snap from prompts
Analyze the preview data from Snaps of a successfully validated pipeline and surface insights into key fields, business and validation rules, and Personally Identifiable Information (PII). These insights are displayed in the SnapGPT chat panel and are automatically used as context when you generate or refine pipelines, resulting in more accurate outputs.
Analyze and describe existing pipelines
Answer questions about the SnapLogic Platform
Interprets requests in the context of what’s already on the canvas and confirms your intention as it processes your input. You can ask for updates using natural language without describing the pipeline in technical terms.
It can deep think, understand the endpoint and suggest a more scalable pipeline for the end users.
Choose from three modes to control the level of AI reasoning for your task:
- Standard: The default mode. Quick responses for all SnapGPT skills.
- Think: Deeper AI reasoning for analysis, Q&A, descriptions, and expression or SQL generation.
- Plan: Conversational planning loop for pipeline generation and refinement.
Learn more about SnapGPT modes.

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.

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.