read_image — load an image as base64

read_image returns an image as base64 plus its media type, ready to feed into a multimodal Anthropic / OpenAI / Google request.

Schema

Arg	Type	Required	Description
path	string	yes	Local file path or http:// / https:// URL.

Example invocation

{"path": "screenshots/dashboard.png"}

from chimera.tools.image_read import ReadImageTool

tool = ReadImageTool()
result = tool.execute(
    {"path": "https://example.com/diagram.png"},
    env=local_env,
)

print(result.metadata["media_type"])  # "image/png"
print(len(result.metadata["image_data"]))  # base64 string length

Output sample

Read 87.2 KB image: screenshots/dashboard.png (image/png)

result.metadata carries:

Key	Value
image_data	base64 string
media_type	e.g. image/png, image/jpeg
path	echo of the input path

Supported formats

PNG, JPEG, GIF, and WebP — the formats accepted by every major multimodal provider.

See also

• browser — screenshot capture from a live browser.
• web_fetch — text fetch.