v2.6.5

Agno now supports multimodal inputs in the Gemini File Search API, so agents can index and semantically search across images alongside text rather than text alone. This is available with google-genai≥1.75.0; older versions remain text-only.

Action required: Bump google-genai to 1.75.0 or later to use image inputs with Gemini file search. Existing text-only setups keep working without changes.

See the cookbook example for an image-upload walkthrough.