Migrate to Extend - Chunkr

The full guide for moving to Extend lives in the current documentation. It applies to v1 integrations as well: the v1 TaskResponse output (chunks containing segments with segment_type, bbox, and content) has the same structure as the current API, so the Parse and Extract mappings carry over directly.

Migration overview

Step-by-step migration plan.

Concept mapping

Every Chunkr concept and its Extend equivalent.

Migrating Parse

Segments to blocks, bounding boxes, chunking, spreadsheets.

Migrating Extract

Schema rules, citations, and confidence.

v1-specific notes

Chunkr v1	Extend
`chunkr.upload(file)`	`client.parse_runs.create_and_poll(file={...})`
`chunkr.get_task(task_id)`	`client.parse_runs.retrieve(id)`
`task.poll()`	Built into `create_and_poll`
`task.update(config)`	Create a new run with the new `config`
`task.cancel()` / `task.delete()`	`client.parse_runs.cancel(id)` / `client.parse_runs.delete(id)`
`task.html()` / `task.markdown()` / `task.content()`	Join `chunk.content` across `run.output.chunks`
`Configuration(...)` object	`config` dictionary on the request
`pipeline`, `segmentation_strategy`, `ocr_strategy`	See the configuration mapping
Self-hosting (Docker Compose, Kubernetes)	Extend offers BYOC and Hybrid deployments; see the FAQ