01
Turn a data request into a pipeline
Tell your assistant what you want to move and where it should go. It can inspect your connections, explore source objects, and prepare the pipeline for review.
Agentic data pipelines
Put your AI assistant to work on data pipelines
Tell Supaflow what data you need and where it should go. Your assistant can explore your connections, prepare a pipeline, run a sync, investigate failures, and manage schedules—all in ChatGPT or Claude.
Works with your Supaflow account, existing connections, and workspace permissions.
From request to pipeline plan
Illustrative workflowYou
Plan a Salesforce-to-Snowflake pipeline for Accounts and Opportunities. Show me the plan before creating it.
I’ll check your workspace, connections, and Salesforce catalog first.
Source: Salesforce Destination: Snowflake Selected objects: Account, Opportunity
Here is the pipeline plan. Would you like me to create it with these objects?
Use the connections you already have
Keep working with your Supaflow sources, destinations, and pipelines.
Carry the task through multiple steps
Move from discovery to a plan, or from a requested run to its job outcome.
Keep decisions in view
Review pipeline scope and get explicit confirmation for destructive actions.
More than a pipeline suggestion
Hand over the task. Keep the context.
Pipeline work often means finding a connection, checking objects, setting up a run, and interpreting its result. Give your assistant the task and let it work through those steps with Supaflow.
02
Run a sync and follow the result
Ask for fresh data from an existing pipeline. Your assistant can start the run, check its progress, and summarize the outcome reported by Supaflow.
03
Get to the next step after a failure
Bring job details and available logs into the conversation. Understand the reported error and review the next action before changing your setup.
04
Change the objects you move
Inspect a pipeline’s current selection, explore the source catalog, and update the scope when your reporting needs change.
05
Make working pipelines repeatable
Set a cadence and timezone, review the schedule, and manage recurring runs once your pipeline is ready.
06
Understand the workspace you already have
Find your connections and pipelines, inspect their configuration, and check recent jobs without assembling the context yourself.
See the work in conversation
Ask for an outcome, then work through the details
These illustrative conversations show supported workflows. Names, responses, and progress are examples; your assistant uses the resources and results returned from your own workspace.
Workflow 1
Describe the pipeline. Review the plan.
Start with the data you need. Your assistant can look at the available source objects and destination, then prepare a concrete plan.
Choose the source, destination, and object scope before creating the pipeline.
Describe the pipeline. Review the plan.
Illustrative workflowYou
Plan a Salesforce-to-Snowflake pipeline for Accounts and Opportunities. Show me the plan before creating it.
I’ll check your workspace, connections, and Salesforce catalog first.
Source: Salesforce Destination: Snowflake Selected objects: Account, Opportunity
Here is the pipeline plan. Would you like me to create it with these objects?
Workflow 2
Ask for a sync. Stay with the job.
Use the pipeline you already have. Follow the run through its actual job status, then see the reported outcome in the conversation.
Follow a requested run from its initial status through to the reported outcome.
Ask for a sync. Stay with the job.
Illustrative workflowYou
Run my Salesforce-to-Snowflake pipeline and monitor it to completion.
I’ll verify the pipeline and start a normal sync using its current object selection.
Example job progression: queued → running → completed
The job is complete. I’ll summarize the result and any metrics returned by Supaflow.
Workflow 3
Investigate a failure with the evidence in view.
Ask about the failed job instead of searching through dashboard pages. Your assistant can inspect reported errors and explain what needs attention.
Review the diagnosis and recommended action before changing a connection or running the pipeline again.
Investigate a failure with the evidence in view.
Illustrative workflowYou
Why did my latest pipeline run fail? Explain the error and recommend the next action.
I’ll inspect the failed job and the available details before recommending a change.
Example error: the destination connection needs attention.
Check the destination connection in Supaflow first. Once it is working, we can review whether to run the pipeline again.
Workflow 4
Turn a working sync into a routine.
Describe when the pipeline should run. Your assistant can inspect existing schedules and prepare the requested cadence and timezone for review.
Make the schedule explicit so recurring runs match the way your team works.
Turn a working sync into a routine.
Illustrative workflowYou
Schedule this pipeline every weekday at 7 a.m. America/Los_Angeles. Show me the schedule before saving it.
I’ll check the pipeline’s existing schedules, then prepare the requested timing.
Cadence: Monday–Friday Time: 07:00 Timezone: America/Los_Angeles
Would you like me to save this schedule for the selected pipeline?
Your assistant reasons. Supaflow runs the pipeline.
The conversation stays grounded in the platform you use for data movement. Your assistant works with Supaflow’s connections, source catalogs, pipeline configuration, and job results.
Real workspace context
The assistant can inspect accessible resources before planning a change. It does not need to guess what connections or objects exist.
Visible scope and permissions
Review the source, destination, and selected objects. Account permissions apply, and destructive operations require explicit confirmation.
Results you can follow
Supaflow executes the requested work. Your assistant can check job status, inspect available details, and summarize the outcome it receives.
Start where you already work
Bring Supaflow into your AI assistant
Choose your assistant and follow its guide to install Supaflow, connect your account, and start with your workspace.
Supaflow for ChatGPT
Install the Supaflow plugin from the public ChatGPT listing. Plan a pipeline, follow a sync, or investigate a failed job in chat.
Use Supaflow in ChatGPTSupaflow for Claude
Connect Supaflow to Claude. Explore your available connections and source objects, then work through your next pipeline task.
Use Supaflow in ClaudeStart with the data you already need
New to Supaflow? Create an account and configure your source and destination connections. Already using Supaflow? Connect your existing account and ask your assistant to inspect a pipeline or prepare your next one.
Supaflow service pricing and your AI client’s access requirements apply. See Supaflow pricing.
Questions before you connect
What are agentic data pipelines?
Agentic data pipelines use an AI agent to carry out multiple steps in pipeline work, such as inspecting connections, planning a pipeline, starting a run, and checking its outcome. With Supaflow, you request the task in your assistant while Supaflow provides the pipeline resources and execution.
What can I do with Supaflow today?
Explore connections and source catalogs, prepare and create pipelines, change object selections, run syncs, inspect jobs and available logs, and manage schedules. Access follows the connected account’s workspace permissions.
Is this agentic ETL or agentic data engineering?
These terms describe a broad category of AI-driven data work. Supaflow focuses on practical data integration workflows: building and operating pipelines through conversation. The examples on this page show the current scope rather than promising every data engineering task.
Does the assistant continuously monitor and automatically repair my pipelines?
The workflows shown here run when you request them in your assistant. You can ask it to monitor a job, investigate failures, or review a change. This page does not describe an unattended monitoring or automatic repair service.
Do I need a Supaflow account and connections?
Yes. Connect your Supaflow account and select a workspace. Configure source and destination connections in Supaflow before creating or running a pipeline. Existing users can work with the connections and pipelines they already have.
Can I deploy a private agent or build new connectors?
The client guides cover separate Supaflow Code editions for local Codex and Claude Code. Install the matching edition from GitHub for private-agent workflows with a local shell and the Supaflow CLI. Hosted chat plugins cannot manage local agents. Guided development and testing of new connectors are coming soon.