Documents

The Python SDK exposes the canonical document viewer through client.documents. Use it when you need full parsed text for a preview, editor, or custom document UI.

Attachments

from agentflow import AsyncAgentFlow

async with AsyncAgentFlow.from_profile("local") as client:
    document = await client.documents.attachment("file_abc123")
    print(document.filename)
    print(document.content)

For agent-scoped attachment checks, pass agent:

document = await client.documents.attachment("file_abc123", agent="agent_abc123")

client.attachments.content(file_id, agent=...) returns the same model and uses the canonical viewer endpoint under the hood.

Knowledge Base Documents

Prefer document_id from client.knowledge_bases.list_files(...) or client.knowledge_bases.list_documents(...).

files = await client.knowledge_bases.list_files("kb_abc123")
document = await client.documents.knowledge_base(
    "kb_abc123",
    document_id=files[0].document_id,
)

print(document.file_type)
print(document.content)

You can also call the KB convenience method:

document = await client.knowledge_bases.view_document(
    "kb_abc123",
    document_id="doc_abc123",
)

Use file_path only when the caller has a source path but no document ID:

document = await client.documents.knowledge_base(
    "kb_abc123",
    file_path="handbook.md",
)

Response Model

All viewer helpers return DocumentContent.

Field	Type	Description
`filename`	`str`	Display filename
`file_path`	`str \| None`	Source path or viewer path
`content`	`str`	Full parsed text
`size`	`int`	Source file size in bytes when available
`last_modified`	`datetime \| None`	Last modified timestamp when available
`file_type`	`str \| None`	File extension/type label
`encoding`	`str \| None`	Text encoding
`summary`	`str \| None`	Attachment summary when available

​Documents

​Attachments

​Knowledge Base Documents

​Response Model

Documents

Attachments

Knowledge Base Documents

Response Model