Images from tools

Some tools produce images for the user — a CAD screenshot, a plotted result. They ride the ToolCall’s finish event as event.images, a tuple[str, ...] of image data: URIs (empty on the start event, and from apps too old to send them). Decode each on the finish event:

import base64, re
from pathlib import Path
from cosmon_agent_sdk import Client, ToolCall

with Client() as client:
    for i, event in enumerate(client.run("Screenshot the open part", agent="solidworks")):
        if isinstance(event, ToolCall) and event.finished:
            for uri in event.images:                       # () when the tool produced none
                m = re.match(r"^data:image/\w+;base64,(.+)$", uri)
                if m:                                       # http links stay stream-only
                    Path(f"shot-{i}.png").write_bytes(base64.b64decode(m[1]))

Each data: image is also embedded in the run’s final result as a spec-native MCP image content block, so plain MCP clients see it too. .text() is unaffected — it joins only the text.

Only data: URIs are embedded in the final result — a tool that returns an http(s) image link surfaces it on the stream event but can’t embed it, so it stays stream-only.

Reference: ToolCall.images

The images field on the streamed ToolCall event.

Unattended batch Multi-agent handoff

Getting Started

CAD

CAE

Agent SDK

Deployment

Reference: ToolCall.images