FoundationAgents/OpenManus · error · ValueError

Unsupported image format: {image}

Error message

Unsupported image format: {image}

What it means

Raised by LLM.ask_with_images() when an element of the images argument is neither a string (treated as a URL or data-URI), nor a dict with a "url" key, nor a dict with an "image_url" key. Any other type — bytes, PIL objects, numpy arrays, file paths to local files without a data: scheme — is rejected because it cannot be mapped to the OpenAI image_url content-part shape.

Source

Thrown at app/llm.py:556

                [{"type": "text", "text": content}]
                if isinstance(content, str)
                else content
                if isinstance(content, list)
                else []
            )

            # Add images to content
            for image in images:
                if isinstance(image, str):
                    multimodal_content.append(
                        {"type": "image_url", "image_url": {"url": image}}
                    )
                elif isinstance(image, dict) and "url" in image:
                    multimodal_content.append({"type": "image_url", "image_url": image})
                elif isinstance(image, dict) and "image_url" in image:
                    multimodal_content.append(image)
                else:
                    raise ValueError(f"Unsupported image format: {image}")

            # Update the message with multimodal content
            last_message["content"] = multimodal_content

            # Add system messages if provided
            if system_msgs:
                all_messages = (
                    self.format_messages(system_msgs, supports_images=True)
                    + formatted_messages
                )
            else:
                all_messages = formatted_messages

            # Calculate tokens and check limits
            input_tokens = self.count_message_tokens(all_messages)
            if not self.check_token_limit(input_tokens):
                raise TokenLimitExceeded(self.get_limit_error_message(input_tokens))

View on GitHub (pinned to 52a13f2a57)

Solutions

1. Convert local images to a data URI string: "data:image/png;base64," + base64.b64encode(open(p,"rb").read()).decode()
2. Pass dicts in the exact shapes: {"url": "..."} or {"image_url": {"url": "..."}}
3. For hosted images, pass the public https URL string directly

Example fix

# before
images = [open("chart.png", "rb").read()]  # bytes -> ValueError

# after
import base64
images = ["data:image/png;base64," + base64.b64encode(open("chart.png", "rb").read()).decode()]

def is_valid_image(image: object) -> bool:
    if isinstance(image, str):
        return True
    if isinstance(image, dict):
        return "url" in image or "image_url" in image
    return False

from typing import TypeGuard, Any

def is_supported_image(image: Any) -> TypeGuard[str | dict]:
    return (
        isinstance(image, str)
        or (isinstance(image, dict) and ("url" in image or "image_url" in image))
    )

When it happens

Trigger: Passing images=[open("x.png","rb").read()] (bytes), images=[PIL.Image.open(...)], images=[{"src": "data:image/png;base64,..."}] (wrong key name), or a plain local path string that is fine syntactically but will fail at the API — only the three accepted shapes pass the guard.

Common situations: Loading images as bytes or PIL objects and assuming the client encodes them; using "src" or "data" keys instead of "url"; forgetting to base64-encode local files into a data URI string.

Related errors

• Model {self.model} does not support images. Use a model from
• The last message must be from the user to attach images
• Invalid tool_choice: {tool_choice}
• Server URL is required for SSE connection
• Command is required for stdio connection

AI-assisted analysis of FoundationAgents/OpenManus@52a13f2a57 (2026-08-15). Data as JSON: /api/errors/c2ed0a4b44646f9a. Report an issue: GitHub.