Separate Documents

ai.segment: Separate a concatenated batch (one big scan) into typed document instances using an LLM. Consumes ai.parse output and a type taxonomy; discovers an unknown number of documents in any order and emits per-document page ranges + text + type, ready for type-specific ai.extract via control.parallel_map. The inverse of ai.split.

Configuration

Configuration goes inside the step’s with: block.

string

required

Template expression resolving to ai.parse output, e.g. ”{{ steps.parse.output }}”

array<object>

required

The document-type taxonomy. The LLM tags each detected document with one of these names, or the reserved “unknown” type when none fit.

Show documentTypes properties

string

required

Type label returned in documents[].type (e.g. “timesheet”, “deal-memo”)

string

What this document type looks like, headings, layout, vocabulary, and the cues that mark its FIRST page. Fed to the LLM for both boundary detection and typing.

string

Optional natural-language rules appended to the system prompt.

string

Provider ID from eigenpal.config.yaml. Falls back to the tenant default LLM provider when omitted.

integer

Override the per-window token ceiling. Defaults to env SPLIT_WINDOW_TOKEN_BUDGET or 20000.

Output

Batch-level escalation summary: labelled vs unlabelled counts + docs needing review.

Show summary properties

Total documents discovered.

Documents assigned a known taxonomy type.

Documents the segmenter could not label (type “unknown”), escalate to a human.

Documents with low boundary/type confidence.

Per-type document counts.

Documents to escalate to a human operator, with the reason for each.

Show needsReview properties

Documents discovered in the batch, in page order. The full batch is covered.

Show documents properties

0-based position of this document in the batch

Chosen type name from the taxonomy, or “unknown”

[startIndex, endIndex] inclusive, 0-based page indices

LLM confidence at the boundary + type. Coarse enum (low | medium | high).

LLM justification for the boundary, useful for debugging

Structured start evidence, parse this instead of regexing over notes.

Show evidence properties

Verbatim heading/title text the LLM cited as the document start

Page where the start appears

Joined per-page content for this document, ready for downstream ai.extract

Raw per-page records covered by this document, in page order.

Show pages properties

Pages with direct LLM boundary evidence

Pages assigned by deterministic continuity fill

Get started

Concepts

Workflow steps

Guides & tutorials

Changelog

Configuration

Output

​Configuration

​Output

Configuration

Output