The available blocks depend on the loaded models (response cached for 10 s). A block or test missing from this response must not be used. No authentication required.
/pipelines/capabilities
Request body
The following parameters can be included in the request body:
Parameters
Successful response
The following fields are returned in a successful response:
Response fields
blocks
Blocks that can be used right now, for ocr and for audio.
condition_tests
Possible values of condition.params.test.
condition_labels
Layout labels covered by each test.
layout_labels
Labels that layout detection can produce.
limits
Effective limits (nodes, links, pages, parallel regions).
model_versions
Version of each model (null if not provided).
stream
The SSE stream is available.
resume
Resuming an interrupted stream (false: not available).
word_boxes
Word boxes: values (all, low), methods (9, 10), default method and blocks concerned (layout, layout_line).
line_boxes
Line boxes of the layout block: values (all), methods (9, 10, same as word_boxes), default method and blocks concerned (layout).
writing_detect
handwritten_detect block: blocks it goes after and model used (alpha-writing-detect).
writing_blocks
Layout blocks after which handwritten_detect is available.
forbidden_after
Rejected sequences (422 forbidden_block_sequence): model_max downstream of layout_word.
documents
Routes /documents: classify and split, page_texts input, "none" label, 255 labels max, default split threshold 50.
words_writing_type
Route /pipelines/words/writing_type: max_words (20,000).
Examples
Code examples for using this endpoint:
curl https://api-endpoints.alphaedge-ai.com/pipelines/capabilities
Response
Sample API response:
{
"schema_version": 1,
"blocks": {
"ocr": ["layout", "layout_word", "layout_line", "handwritten_detect", "model_max", "model_fast", "model_handwritten", "condition", "confidence"],
"audio": ["diarization", "transcription", "confidence", "postcorrect"]
},
"condition_tests": ["table", "image", "text", "header", "footer", "aside", "handwritten", "printed"],
"layout_labels": ["abstract", "algorithm", "aside_text", "chart", "…", "table", "text", "vertical_text", "vision_footnote"],
"condition_labels": {
"table": ["table"],
"image": ["image", "header_image", "footer_image"],
"text": ["text", "abstract", "content", "reference", "reference_content", "vertical_text"],
"header": ["header"], "footer": ["footer"], "aside": ["aside_text"],
"handwritten": ["handwritten", "mixed"], "printed": ["printed", "unknown"]
},
"limits": { "max_nodes": 40, "max_links": 80, "max_pages": 100, "max_concurrent_regions": 8 },
"model_versions": { "alpha-layout": null, "alpha-digit-max": null, "alpha-audio-v1": null, "…": null },
"stream": true,
"resume": false,
"word_boxes": { "values": ["all", "low"], "methods": ["9", "10"], "default_method": "9", "blocks": ["layout", "layout_line"] },
"line_boxes": { "values": ["all"], "methods": ["9", "10"], "default_method": "9", "blocks": ["layout"] },
"writing_detect": { "block": "handwritten_detect", "after": ["layout", "layout_line", "layout_word"], "model_slug": "alpha-writing-detect" },
"writing_blocks": ["layout", "layout_line", "layout_word"],
"forbidden_after": { "layout_word": ["model_max"] },
"documents": { "routes": ["classify", "split"], "input": "page_texts", "none_label": "none", "max_labels": 255, "split_default_threshold": 50 },
"words_writing_type": { "max_words": 20000 }
}
Error
503 model_unavailable if the AI server does not respond.