> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://api.docs.siftsy.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://api.docs.siftsy.com/_mcp/server.

# BI sync

> Export key analyses and insights into your data warehouse or data platform: a suggested load order for Snowflake, BigQuery, Databricks, Redshift, and BI tools.

The export side of the Siftsy Analyst API. A typical full sync into a warehouse or BI model:

#### Collections dimension

`GET /collections` — build the collection table. Campaigns include nested groups; posts join via `collectionIds`.

#### Posts fact table

`GET /posts?limit=100` and follow `nextCursor`. Default status is `complete`. Use `status=all` only if you need pending/archived/unreachable rows.

#### Per-post metrics and analysis

`GET /posts/{id}/metrics` — daily score rollups and mix buckets. Skip or retry 404 until scoring completes.
`GET /posts/{id}/analysis` — content + comment analysis.
Skip a per-post comments crawl. If you need evidence for a specific analysis, sample it with `GET /posts/{id}/analysis/comments?source=notables` — not a full-thread export.

#### Campaign analyses

`GET /analyses?kind=report` — list persisted campaign reports (up to 1,000 posts each).
`GET /analyses/{id}` — the report. Comment bodies are not included.
`GET /analyses/{id}/comments?source=notables` — a sample of comments behind that analysis, not the full sections.

#### Rollups per collection or tag (optional)

`POST /analyses` with `selection.collectionId` or `selection.posts.tag` — one aggregate per campaign / tag, computed from stored scores. Selections of ≤ 100 posts return `201 ready` inline; larger ones return `202` and are polled at `GET /analyses/{id}`. Recreate on each sync — analyses expire after 7 days. See [Analyses](/guides/core-concepts/analyses).

## Refresh schedule

A full sync of a 500-post workspace is about 1,070 requests and 15 MB. Budgets fit eight of them a day (the Power BI Pro scheduler maximum), or — cheaper — one full sync plus hourly incremental runs (`GET /posts?updated_since=<last run>`, then metrics and analysis only for the posts returned). Recreate rollups on one sync a day, not every run. Run the refresh queries sequentially or at most two in parallel; the per-minute caps are sized so a sequential loop never sees a `429`. See [Rate limits](/guides/core-concepts/rate-limits) for the worked budget.

## Data classification

Comment text and usernames are customer data. `addedByEmail` identifies the teammate who added the post (for automation-added posts, the teammate who set up the automation); `addedVia` is `api` for posts added through the [content routes](/guides/core-concepts/adding-content), `automation` for posts a Siftsy Automation pulled in on a schedule, and `dashboard` otherwise. `automation` (`null` or `{ addedById, ids, lastRefreshAt, refreshCount }`) lists every automation that has added or refreshed the post, so a manually added post that an automation keeps refreshed shows `addedVia: dashboard` with a non-empty `automation.ids`. Recipients are responsible for retention and downstream sharing.

Demo or admin-uploaded posts may appear with `isDemo: true` and do not imply billing usage.