browser-use/browser-use · error · ValueError

Failed to download image from {url}: {e}

Error message

Failed to download image from {url}: {e}

What it means

Raised when downloading an http(s) image for Bedrock fails: the httpx GET raised, returned a non-2xx status (raise_for_status), or any other exception occurred during fetch. The original error is embedded in the message.

Source

Thrown at browser_use/llm/aws/serializer.py:88

			response.raise_for_status()

			# Detect format from content type or URL
			content_type = response.headers.get('content-type', '').lower()
			if 'jpeg' in content_type or url.lower().endswith(('.jpg', '.jpeg')):
				image_format = 'jpeg'
			elif 'png' in content_type or url.lower().endswith('.png'):
				image_format = 'png'
			elif 'gif' in content_type or url.lower().endswith('.gif'):
				image_format = 'gif'
			elif 'webp' in content_type or url.lower().endswith('.webp'):
				image_format = 'webp'
			else:
				image_format = 'jpeg'  # Default format

			return image_format, response.content

		except Exception as e:
			raise ValueError(f'Failed to download image from {url}: {e}')

	@staticmethod
	def _serialize_content_part_text(part: ContentPartTextParam) -> dict[str, Any]:
		"""Convert a text content part to AWS Bedrock format."""
		return {'text': part.text}

	@staticmethod
	def _serialize_content_part_image(part: ContentPartImageParam) -> dict[str, Any]:
		"""Convert an image content part to AWS Bedrock format."""
		url = part.image_url.url

		if AWSBedrockMessageSerializer._is_base64_image(url):
			# Handle base64 encoded images
			image_format, image_bytes = AWSBedrockMessageSerializer._parse_base64_url(url)
		elif AWSBedrockMessageSerializer._is_url_image(url):
			# Download and convert URL images
			image_format, image_bytes = AWSBedrockMessageSerializer._download_and_convert_image(url)
		else:

View on GitHub (pinned to 6c73fced2f)

Solutions

1. Verify the URL is reachable (curl -I) from the machine running the agent
2. If the URL is presigned/expiring, refresh it or download the image yourself and pass a base64 data URL
3. Increase reliability by hosting images somewhere stable and public to the agent

Example fix

```python
# before - rely on remote URL
img = {'type': 'image_url', 'image_url': {'url': 'https://cdn/expiring.png'}}

# after - fetch once yourself and inline
import httpx, base64
b = httpx.get(url, timeout=30).content
img = {'type': 'image_url', 'image_url': {'url': 'data:image/png;base64,' + base64.b64encode(b).decode()}}
```

```python
import httpx
def url_fetchable(url: str) -> bool:
    try:
        return httpx.head(url, timeout=10, follow_redirects=True).status_code < 400
    except Exception:
        return False
```

```python
try:
    await llm.ainvoke(msgs)
except ValueError as e:
    if 'Failed to download image' in str(e):
        b64 = inline_image_as_data_url(url)  # fetch yourself, send base64
```

When it happens

Trigger: 404/403 on the image URL, expired signed URLs (presigned S3 links), DNS failure, TLS errors, server timeouts beyond the 30s limit, or a content-type the code cannot classify (it defaults to jpeg but the fetch itself may still fail).

Common situations: Screenshots hosted on short-lived CDNs; internal URLs not reachable from the agent's network; hotlink protection blocking httpx's default headers.

Related errors

• {e.message}
• Invalid base64 URL: {url}
• Failed to decode base64 image data: {e}
• httpx not available. Please install it to use URL images wit
• Unsupported image URL format: {url}

AI-assisted analysis of browser-use/browser-use@6c73fced2f (2026-08-14). Data as JSON: /api/errors/d6fd7fddde8560e7. Report an issue: GitHub.