HarnessAgent now supports image inputs for all harness adapters
Vercel's AI SDK layer now accepts image inputs across every HarnessAgent harness adapter. Images can be passed alongside text in the standard AI SDK message format, using a content array that mixes text and file parts.
Why it matters
Builders using HarnessAgent can send images with text prompts to any supported harness without adapter-specific code. Image input is processed by each harness's underlying vision model, so output quality depends on those models.
Who should care
Developers building AI agents on Vercel's AI SDK, teams using HarnessAgent with multiple harness adapters, and builders of image-aware workflows such as design review or product photo tooling.
What you can do
Read the HarnessAgent image inputs documentation and send a test message with a content array containing a text part and a file part to one of your configured adapters.