Slack
warehouse
Read-only by design
Writes are refused at the database, not left to the model.
Discovery
Better answers start with understanding your data.
A field called 'available' doesn't always mean a customer can book it. Refix examines the data behind the names, so that distinction shapes the investigation from the start.
Used by teams at
Your data has its own language. Refix learns it.
-
Find the relevant data
Connect a source and Refix explores its structure, building a map it can use in later investigations.
-
Understand what the fields mean
Actual values and relationships give meaning to unfamiliar names, so the answer starts from the right data.
-
Connect the same story across tools
Refix identifies likely links between records and notes its confidence, giving investigations a path across systems.
Tables and fields
A column name only tells part of the story.
A status code needs a meaning before it can support an answer. Refix examines sample values and the surrounding schema to learn what a table represents and how to use it.
- Descriptions grounded in the structure and values
- Keys and relationships that guide queries
- Filters out likely backup and temporary tables
Across tools
Follow one customer across the systems they touch.
A booking ID can connect a product event to a payment. Refix identifies those relationships, so an investigation can follow what happened beyond the boundaries of one tool.
- Identifies shared fields across sources
- Records confidence in each proposed link
- Works within the access you provide
More of Refix
See how Refix finds an opportunity and follows it through.
- Watchers Growth goals followed through
- Sessions Work through a business question
- Investigations Find the cause and the next step
- Memory Builds on what your team has learned
- Charts Visual explanations your team can share
- Slack Work with Refix in the conversation
- Integrations Find opportunities across your tools
Security and control, built in.
Choose what Refix can access and set the approvals it needs. Review the data and reasoning behind its recommendations.
Approvals you control
Require your team's approval before Refix takes action in connected tools.
Roles for every teammate
Decide who connects data, approves actions, and manages billing.
Usage you can see
Daily credit use, with an alert before you reach your limit.
Zero data retention
Model providers keep nothing and never train on your data.
Self-hosting
Understand your data inside your own cloud.
Run the Refix agent on your own Kubernetes cluster. Your platform team controls access and model connections, with our engineers helping you deploy.
- Helm chart for your Kubernetes cluster
- Your model keys or LLM gateway
- Slack sign-in or your identity provider
- Network access controlled by your firewall
- Application updates you approve
- Controls for reporting back to Refix
What does Discovery do?
Discovery builds a map of your connected data and what it means. It examines structures and sample values, then records relationships that help Refix choose the right data for an investigation.
Which sources does Discovery map?
Databases and warehouses like Postgres and BigQuery, BI tools like Metabase, and product analytics like PostHog, Mixpanel, and Google Analytics.
How does Discovery link data across tools?
It looks for fields that identify the same entity across sources, such as a customer or booking ID. Each proposed relationship has a confidence level that informs later investigations.
Does Discovery define our metrics?
Your team's definitions guide the analysis. Discovery learns the structure and meaning of your data; Memory keeps the business context you provide, including how your team defines a metric.
Do we need to prepare our data first?
Connect a supported source with the right access, and Refix starts exploring it. Existing definitions and documentation help it learn. It filters out likely backups and temporary tables during discovery.