Lantide Data
← Back to blog
Integrate with Lantide

Why Do You Still Need Lantide When Using Hex?

Hex excels at collaborative cloud notebooks, reactive execution, and app publishing. When data must stay local or ad hoc exploration does not deserve a cloud project, use Lantide for local-first analysis and send sanitized results back to Hex.

Hex is well suited to cloud collaboration: projects are made of cells, a reactive graph rebuilds dependent outputs, and teams get Graph View, kernels, and publishable Apps. See Execution model and Project kernels.

If the data is approved for the cloud and several teams need to edit and view it, keep using Hex. Lantide covers a different situation: sensitive files must stay local, or an Excel file that arrived today is not worth turning into a complete cloud project. Get the definition right locally first, then send a sanitized result to Hex for publication.

When to start locally

1. The file cannot enter a cloud kernel yet

Financial, healthcare, and highly regulated teams often prohibit unredacted local reports, customer files, or internal credentials from entering a third-party compute environment. Hex has security and AI privacy controls, including Enterprise BYOK; when organizational policy still says “clean locally before sharing,” a local-first workstation is needed. See Credits and AI data privacy.

2. Temporary exploration should not require a full cloud project

A spreadsheet that arrives today and may be discarded tomorrow can make a kernel and collaboration workflow unnecessarily heavy. A local IDE separates exploration cost from the formal publishing environment.

3. Your definition of done is different

Hex is strong at publishable data Apps and collaborative canvases. Lantide is strong at SQL-first Plan review, annotations, persistent SQL, Source Run, and Report evidence: one favors interactive delivery, the other a traceable analysis contract.

Stage Where Why
Sensitive CSV / local-file exploration Lantide Query files directly with local DuckDB
Definition alignment and formal SQL Lantide Plan, annotations, Approve & Execute, SQL evidence
Sanitized results for cross-team use Hex Collaboration, App publishing, internal sharing
Long-term monitoring dashboard Hex or existing BI Lantide is not a shared monitoring platform

Querying locally does not mean the entire path sends nothing outside the device. If BYOK connects to OpenAI or Anthropic, prompts and necessary context may still be sent; inspect external databases and MCP sources individually. See Local-first AI data analysis.

What the local stage gives you

DuckDB and SQL tabs

CSV, Parquet, and Excel files can be queried directly. Intermediate results can be materialized as paginated caches, and Source Run can rerun the dependency chain. See Query Excel / CSV with DuckDB and USER_GUIDE §7–8.

Project Focus and Compare View

Enable Project Focus in the sidebar so the Agent stays focused on that project's Plan and Report. When comparing an old definition, open the Plan or SQL in Compare View while continuing to edit in the main window. See USER_GUIDE §11.4 and §11.10.

BYOK for the built-in Agent

Lantide's built-in Agent uses your own model key and defaults to execution in the local workspace. Hex Enterprise BYOK addresses key and billing ownership in a cloud project; the two can coexist, but their default compute locations differ.

A practical three-step workflow

  1. Explore sensitive or temporary data in Lantide: load local files, draft a Plan with Project Focus, and use Compare View to compare old SQL when needed.
  2. Finish sanitization and definition review: export only approved aggregates or reviewed SQL.
  3. Publish the collaborative result in Hex: bring validated logic or sanitized data into Hex as an interactive team App or document.
Dimension Hex primarily Hex with Lantide
Default compute location Cloud project kernel Local DuckDB
Collaboration and publishing Strong Mainly export / handoff
Sensitive local files Often policy-limited Good for local validation first
AI billing Credits; Enterprise can use BYOK Built-in Agent BYOK; provider fees remain your responsibility

Conclusion

Hex turns analysis into a shareable cloud product; Lantide helps you get right, locally, the work that cannot or should not go to the cloud yet. Local validation followed by cloud publishing is often closer to a real workflow than choosing only one.

Next steps

  • Read Local-first AI data analysis.
  • Take a spreadsheet that cannot currently enter the cloud, validate its definitions in Lantide, and then decide whether to sanitize and import it into Hex.

References