mirror of https://github.com/crewAIInc/crewAI.git synced 2026-01-23 15:18:14 +00:00

Files

Greyson LaLonde dc4bbfb5b9 feat(files): add api parameter to format_multimodal_content

Allows selecting OpenAIResponsesFormatter via api="responses" parameter
instead of always using Chat Completions format.

2026-01-23 02:06:09 -05:00

src/crewai_files

feat(files): add api parameter to format_multimodal_content

2026-01-23 02:06:09 -05:00

tests

test: add Gemini video and audio integration tests

2026-01-22 23:16:08 -05:00

pyproject.toml

docs: add README and description for crewai-files package

2026-01-22 22:50:10 -05:00

README.md

docs: add README and description for crewai-files package

2026-01-22 22:50:10 -05:00

README.md

crewai-files

File handling utilities for CrewAI multimodal inputs.

Supported File Types

ImageFile - PNG, JPEG, GIF, WebP
PDFFile - PDF documents
TextFile - Plain text files
AudioFile - MP3, WAV, FLAC, OGG, M4A
VideoFile - MP4, WebM, MOV, AVI

Usage

from crewai_files import File, ImageFile, PDFFile

# Auto-detect file type
file = File(source="document.pdf")  # Resolves to PDFFile

# Or use specific types
image = ImageFile(source="chart.png")
pdf = PDFFile(source="report.pdf")

Passing Files to Crews

crew.kickoff(inputs={
    "files": {"chart": ImageFile(source="chart.png")}
})

Passing Files to Tasks

task = Task(
    description="Analyze the chart",
    expected_output="Analysis",
    agent=agent,
    input_files=[ImageFile(source="chart.png")],
)