The model does not support network images

Error message

The model does not support network images: {network_image_urls:?}

What it means

Identical guard to the Bedrock one, but in the Claude (Anthropic API) client's body builder: claude_build_chat_completions_body collects network image URLs from messages and, because the Anthropic Messages API also requires images as base64 source blocks rather than remote URLs, fails fast with the list of offending URLs.

Solutions

Download each image and supply it as base64-encoded source data in the message content
Preprocess the prompt to inline image bytes before calling the client
Route URL-image workloads to a provider that supports them (e.g. OpenAI vision)
Remove image content if the Claude model does not need it

Example fix

// before
{ "type": "image", "url": "https://example.com/img.jpg" }
// after
let b64 = base64::encode(download("https://example.com/img.jpg").await?);
{ "type": "image", "source": { "type": "base64", "media_type": "image/jpeg", "data": b64 } }

Defensive patterns

Strategy: validation

Validate before calling

for img in message_images {
    if img.is_url() {
        let bytes = download(img.url()).await?;
        img.to_base64(bytes)?; // convert before send
    }
}

Type guard

fn is_url_image(img: &Image) -> bool {
    matches!(img, Image::Url(u) if u.starts_with("http"))
}

Prevention

Base64-encode all images before Anthropic API calls
Preprocess messages in one place so no URL image leaks through
Test prompts containing images against the Anthropic client in CI
Strip images when targeting models/modes that don't need them

When it happens

Trigger: prepare_chat_completions -> claude_build_chat_completions_body encounters user messages containing images referenced by http(s) URL rather than base64-decoded image data, making network_image_urls non-empty.

Common situations: Using image URLs in prompts against an Anthropic-backed client, code migrated from providers that accept URL images, or pipelines that never download/encode the images.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

AI-assisted analysis of sigoden/aichat@82976d349a (2026-09-09). Data as JSON: /api/errors/9b17cd31e86b5db9. Report an issue: GitHub.

Appendix: source

Thrown at src/client/claude.rs:260

                        }));
                    }
                    vec![
                        json!({
                            "role": "assistant",
                            "content": assistant_parts,
                        }),
                        json!({
                            "role": "user",
                            "content": user_parts,
                        }),
                    ]
                }
            }
        })
        .collect();

    if !network_image_urls.is_empty() {
        bail!(
            "The model does not support network images: {:?}",
            network_image_urls
        );
    }

    let mut body = json!({
        "model": model.real_name(),
        "messages": messages,
    });
    if let Some(v) = system_message {
        body["system"] = v.into();
    }
    if let Some(v) = model.max_tokens_param() {
        body["max_tokens"] = v.into();
    }
    if let Some(v) = temperature {
        body["temperature"] = v.into();
    }

View on GitHub (pinned to 82976d349a)