ozongo/dseyes

DsEyes is a lightweight image-to-text vision bridge: an agent or user supplies images and a task prompt, and the CLI returns a vision model’s text for the main model to use. The main runtime is concentrated in one script, so its distinct stages are shown as separate responsibilities sharing that path. The GUI edits the same configuration and can test connectivity. Provider protocols and these workflows are documented; sampled source excerpts are partial, so uncertain internal wiring is omitted.