№ 0431Reddit post
DeepSeek agent borrowing Glimmer's eyes
Wired a Pi agent running DeepSeek-V4-Flash to call a local Muse Glimmer endpoint for vision; DeepSeek screenshotted its canvas animation, asked Glimmer to critique it, and iterated for 30–60 minutes.
PandaBearFred told the non-vision DeepSeek model it could send screenshots to Muse-Glimmer for descriptions. The loop produced a finished parallax car animation, though a second run was not reproducible: DeepSeek decided Glimmer's feedback was unreliable and switched to inspecting frames with PIL.
I asked DeepSeek-V4-Flash to work with Muse-Glimmer for Vision ability in PI agent and it produced this
Same old prompt, just appended a TIP in the end: "Write a single HTML file with a full-page canvas and no libraries. Simulate a realistic side-view of a moving car as the main subject. Keep the car visible in the foreground while the background landscape scrolls continuously to create the feeling that the car is driving forward. Use layered scenery for depth: nearby ground, roadside elements, trees, poles, and distant hills or mountains should move at different speeds for a natural parallax effect. Animate the wheels spinning realistically and add subtle body motion so the car feels connected to the road. Let the environment pass smoothly behind it, with repeating but varied scenery that makes the movement feel believable. Use cinematic lighting and a cohesive sky, such as sunset, dusk, or daylight, to enhance atmosphere. The overall motion should feel calm, immersive, and realistic, with a seamless looping animation. TIPS: You don't have vision abilities so don't try it yourself. If you feel in need of vision ability, you can access http://xxx:8080/v1, model id: Muse-Glimmer for help, it will see the picture, and describe it for you." Then the PI agent started spinning, round a
