Free trial
The signup page offers a free start, and the homepage links to it as the entry point for new users. Plan limits and trial duration are not stated on the homepage.
Learn moreHex.tech
Notebooks, agents and self-serve analytics on one shared data context
Intermediate
What is Hex.tech?
Hex.tech is a developer tool. An AI analytics platform that combines an agentic notebook for Python and SQL, natural language Threads for self-serve questions, and published data apps, all grounded in shared governed context that data teams curate.
Hex is an AI analytics platform that puts three normally separate workflows on a single shared layer of context. Analysts get a multiplayer notebook where Python, SQL and no-code cells sit side by side and an agent can draft, edit and debug the code in place. Business users get Threads, a natural language interface that returns full analyses rather than single numbers, with follow-up questions allowed. Between the two sits a data app builder that publishes finished analysis as something non-technical colleagues can explore and query in conversation. The company's positioning rests on the gap between those modes elsewhere: conversational tools and code-based tools usually read from different context, so the same question yields different answers and someone has to reconcile them by hand. Hex keeps both pointing at the same semantic models, database metadata, endorsed data and AI guides, which data teams curate in Context Studio. Any Thread conversation can be opened as a notebook for inspection, so the agent's reasoning is exposed rather than hidden. In practice the tool suits data teams at companies large enough to have a warehouse and a governance story: analytics engineers, data scientists and the product or operations staff who keep filing requests. Slack is a first-class surface, and Threads can also be started from Claude, Cursor or Codex. The Hex CLI lets people already working in a coding agent publish findings back into Hex. Compared with BI dashboards or standalone AI chat tools, Hex expects a real data stack behind it and rewards setup work on semantic models and endorsed datasets. Privacy terms state that model providers do not train on customer data and that session data falls under the applicable Terms of Service.
How do you use Hex.tech?
1Request a demo or open a free workspace
Book a walkthrough with the sales team to see the platform against your own stack, or start a free workspace directly from the signup page.
2Read the AI overview in the docs
The getting started guide explains what the agents can do, how prompts map to notebook cells, and where human review is expected.
3Connect a data source and ask a first question
Attach a warehouse connection, then type a plain-language question into the ask box to see the agent return a complete analysis rather than a bare number.4Curate context for the agent
Use Context Studio to register semantic models, endorse trusted datasets and write AI guides, which raises the accuracy of every later answer.
5Move into the notebook for deeper work
Open the Thread as a project and continue in the notebook, mixing SQL, Python and no-code cells while the agent edits and debugs alongside you.
6Publish as a data app and share to Slack
Turn the analysis into an explorable app that colleagues can chat with, and route recurring questions through Threads in Slack so answers land where the team talks.
Pros and cons
Pros
- Notebook, conversational and app workflows read from one shared context layer instead of producing competing answersAI
- Agent writes, edits and debugs code with awareness of upstream and downstream dependencies in the notebook flowAI
- Every Thread conversation can be reopened as a notebook, so the agent's reasoning stays inspectableAI
- Slack, Claude, Cursor and Codex surfaces put answers where teams already ask data questionsAI
- Context Studio and semantic models give data teams direct control over what the AI is allowed to trustAI
Cons
- Real value depends on prior governance work: semantic models, endorsed data and AI guides must exist before answers become reliableAI
- No pricing figures appear on the homepage, so budget has to be established through a sales conversationAI
- The product assumes a warehouse and a data team; small companies without that infrastructure will find little to useAI
- Published as a web platform only, with no indication of offline or desktop useAI
- Judging answer accuracy requires enough SQL or Python literacy to read the generated codeAI
How much does Hex.tech cost?
Support
A demo request route, a product FAQ, a privacy and data usage FAQ, and getting started documentation covering AI features.
Learn moreIntegrations
Slack for Threads, plus Claude, Cursor and Codex as places to start conversations. Database metadata, semantic models, external repositories and existing dashboards feed the context layer.
Learn moreFeatures
Agentic notebook for Python, SQL and no-code analysis, with an agent that writes, edits and debugs code and tracks dependencies. Threads for conversational self-serve answers with follow-up questions, openable as notebooks. AI-assisted data app builder with chat over published apps. Context Studio for semantic models, endorsed data and AI guides. Slack, Claude, Cursor and Codex access, plus a CLI for bringing your own agent.
Learn moreFrequently asked questions about Hex.tech
Does Hex.tech offer a free trial?
The signup page offers a free start, and the homepage links to it as the entry point for new users. Plan limits and trial duration are not stated on the homepage.How do you use Hex.tech?
The walkthrough on this page covers 6 steps: 1. Request a demo or open a free workspace 2. Read the AI overview in the docs 3. Connect a data source and ask a first question 4. Curate context for the agent 5. Move into the notebook for deeper work 6. Publish as a data app and share to Slack.What platforms does Hex.tech support?
Hex.tech is available on Web App.What does Hex.tech integrate with?
Slack for Threads, plus Claude, Cursor and Codex as places to start conversations. Database metadata, semantic models, external repositories and existing dashboards feed the context layer.What are the limitations of Hex.tech?
Real value depends on prior governance work: semantic models, endorsed data and AI guides must exist before answers become reliable. No pricing figures appear on the homepage, so budget has to be established through a sales conversation. The product assumes a warehouse and a data team; small companies without that infrastructure will find little to use.
