Hermes Agent failed to transmit the map? Check the auxiliary vision model first

Hermes Agent fails to transmit images, so don't doubt the clipboard. The official Vision documentation is straightforward: images are sent to the model as a base64 content block, so the model itself must support visual input. If the current master model is a plain text model, or if you point the auxiliary vision to an endpoint that does not support images, it will fail.

How to check first

1. Confirm that the image has been successfully pasted, and you can see the attachment prompt in the interface.
2. Change to a provider test that confirms that vision is supported.
3. Check if auxiliary.vision.provider, model, base_url are misconfigured.
4. If you use provider: main, the master model must also be a multimodal model.

The official configuration document also reminds that Codex OAuth supports vision in the auto-detection chain; Custom endpoints must be guaranteed to be compatible.

In a word: the failure of image transmission is mostly not "the picture is not pasted", but the auxiliary vision model does not support the image.

Official open source address: https://github.com/NousResearch/hermes-agent; Official document entry: https://hermes-agent.nousresearch.com/.