Lantide Data
← Back to blog
Integrate with Lantide

Why Do You Still Need Lantide When Using Obsidian?

Obsidian and Dataview excel at local Markdown knowledge and note metadata queries. Lantide adds DuckDB SQL, rerunnable analysis lineage, and governed Agent memory; the two connect through files rather than replacing each other.

Obsidian helps you accumulate shared understanding and context; Lantide turns numbers into reviewable, rerunnable analysis.

Obsidian stores notes as local Markdown, and Dataview can query frontmatter, tags, and tasks. See Dataview and Differences to SQL.

When you need multi-table SQL over external CSV / Parquet, intermediate cache lineage, and approval before AI-extracted definitions are written back, the vault and the analysis workstation remain two different layers. Lantide fills the latter.

Where a note vault struggles with analysis

1. Writing down a number is easy; rerunning it is hard

Writing “Q3 repurchase rate: 23%” in a note is quick. Three months later, recomputing it with the same join and exclusions requires SQL, dependencies, and an execution record. See Reproducible analysis.

2. Dataview serves notes, not an analytical warehouse

DQL is strong at filtering, grouping, and listing notes, but its execution model differs from SQL. Even loading a CSV with dv.io.csv() does not provide persistent SQL tabs, a Source Run DAG, or a formal Report. See the Codeblock reference.

Third-party DuckDB or AI plugins vary by plugin; the capabilities below refer to Lantide itself.

3. AI memory can absorb stale definitions

A general note-taking AI plugin may not know the current workspace schema, Plan approval state, or Queued Knowledge flow. If an obsolete rule goes directly into long-term memory, it can affect later reasoning. See Agent Memory needs approval.

How to connect them: Markdown as an exchange format

There is currently no official two-way plugin. Use Markdown as the bridge:

Obsidian (hypotheses, background, reading notes)
        │  Copy or save as Markdown
        ▼
Lantide (Plan → annotation → Execute → Report / SQL)
        │  Export or copy Markdown / HTML
        ▼
Obsidian Vault (archive traceable conclusions; rerun in Lantide when needed)

Original files and the workspace can remain local. If you use a cloud model, prompts and necessary context may still be sent out. See The practical boundary of local-first.

What Lantide adds

Executable Plan / Report

Bring a hypothesis from Obsidian into a Lantide project, shape it into Plan.md, and produce Report.md after annotation and Approve & Execute. Numbers link back to SQL evidence instead of living only in chat.

Source Run and lineage

When an underlying CSV or upstream tab changes, run Source Run on the relevant SQL tabs to rebuild the dependency chain. Materialized caches have a lifecycle; reopening a workspace may require a rerun. Writing that into your SOP is safer than assuming that saving a result into the vault makes it update forever. See USER_GUIDE §8.

Queued Knowledge

Business rules extracted from a conversation first enter Queued Knowledge. Only after approval are they written into User / Project knowledge files under Rules or Info. Unapproved suggestions are not injected into model context. See USER_GUIDE §15.

A practical four-step workflow

  1. Think in Obsidian: organize background, questions, and hypotheses without hard-coding numbers too early.
  2. Model and execute in Lantide: create a project, draft the Plan, align definitions, and Execute.
  3. Preserve evidence: ensure SQL tabs, Steps, and Report limitations are complete.
  4. Save back to the vault: copy or export the Report summary or full Markdown into Obsidian. When it needs updating, return to Lantide for Source Run and replace the note.
Dimension Obsidian primarily With Lantide
Knowledge and the slip-box Yes —
Note metadata queries Dataview —
Large cross-table SQL Depends on plugins Unified DuckDB query layer
Rerunnable analysis lineage Usually custom-built Source Run / persistent SQL
AI memory governance Depends on plugins Queued Knowledge approval
Handoff Local files Copy / export Markdown

Conclusion

Using Obsidian as a local knowledge base for thinking and archiving, and Lantide as the analysis execution and audit environment, is a stable combination for local-first work. Markdown can move between them; “notes can be queried” and “analysis can be audited” are still different claims.

Next steps

  • Read Local-first AI data analysis.
  • Choose a note in the vault with a hard-coded number, rebuild it as a Plan / Report in Lantide, and save the rerunnable version back to the vault.

References