diegosouzapw/OmniRoute · error

Vision API error

Error message

Vision API error: ${anthropicData.error.message || JSON.stringify(anthropicData.error)}

What it means

Anthropic-specific error path in callVisionModelSingle: the response parsed successfully but its top-level `error` object is present, so the error message (or the JSON-serialized error object when message is absent) is re-thrown. The Anthropic vision endpoint reported an error inside the body even though the HTTP layer succeeded — model name, auth or request-shape problem at the provider.

Solutions

1. Read the embedded Anthropic error message for the exact cause (quota, invalid model)
2. Retry with backoff for overloaded errors

When it happens

Trigger: Thrown at src/lib/guardrails/visionBridgeHelpers.ts:855 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of diegosouzapw/OmniRoute@a179ffed5b (2026-08-25). Data as JSON: /api/errors/489a7b8f17d9a4c1. Report an issue: GitHub.

Appendix: source

Thrown at src/lib/guardrails/visionBridgeHelpers.ts:855

    clearTimeout(timeoutId);

    if (!response.ok) {
      const errorText = await response.text().catch(() => "Unknown error");
      throw new Error(`Vision API error ${response.status}: ${errorText}`);
    }

    const data = await readVisionResponseBody(response);

    if (isAnthropic) {
      // Anthropic response format: { content: [{ type: "text", text: "..." }] }
      const anthropicData = data as {
        content?: Array<{ type?: string; text?: string }>;
        error?: { message?: string };
      };

      if (anthropicData.error) {
        throw new Error(
          `Vision API error: ${anthropicData.error.message || JSON.stringify(anthropicData.error)}`
        );
      }

      const textContent = anthropicData.content?.find((c) => c.type === "text");
      const content = textContent?.text;
      if (!content || typeof content !== "string") {
        throw new Error("Vision API returned empty or invalid response");
      }

      return content.trim();
    } else {
      // OpenAI-compatible response format. Falls back to reasoning_content when
      // content is null — reasoning models (e.g. xiaomi/mimo-v2.5) can exhaust
      // max_tokens on chain-of-thought and return content: null with the full
      // analysis in reasoning_content.
      return extractOpenAICompatibleContent(data);
    }

View on GitHub (pinned to a179ffed5b)