{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreic4sje4ckdrgwnmr7bbvv5jrrah2if2jf5wk67pnhplhe5zmlw6iy",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mk3muv5ya6w2"
},
"path": "/t/chatgpt-connector-returns-empty-for-mcp-tool-results-with-type-image-while-same-tool-works-in-mcp-inspector/1375446#post_5",
"publishedAt": "2026-04-22T12:42:26.000Z",
"site": "https://community.openai.com",
"textContent": "This looks like an integration-surface mismatch rather than an invalid MCP tool result.\n\nMCP image blocks like `{ type: \"image\", data: \"<raw base64>\", mimeType: \"image/png\" }` are valid MCP, which is why MCP Inspector can render them. However, ChatGPT connector/app output should not currently be treated as a guaranteed “tool-returned image becomes model-visible vision input” path.\n\nFor ChatGPT, return model-readable output as `content` text and/or `structuredContent`. If you return only an image block, the visible structured body may appear as `{}`.\n\nRecommended workaround:\n\n- Keep the image block if you want clients that support MCP images to render it.\n\n- Also return a text summary and `structuredContent`.\n\n- If the model needs to analyze the screenshot, have your MCP tool call the Responses API with the screenshot as `input_image`, then return the analysis as text/JSON.\n\nHope this helps!",
"title": "ChatGPT connector returns empty {} for MCP tool results with type: \"image\" while same tool works in MCP Inspector"
}