Ovis-Image Text to Image
Ovis-Image is a 7B text-to-image model specifically optimized for high-quality text rendering in generated images. Designed to operate efficiently under computational constraints, it excels at accurately generating images containing text content.
About
This workflow is a focused text-to-image generator built around the Ovis-Image 7B model, which is optimized to render readable, well-placed text directly inside generated images. The core generation logic is encapsulated in a Subgraph node (5a9f8bf0-ed77-40e3-aca8-c7c1a0bf54e1), keeping the graph clean while exposing the controls you need for prompting and iteration. A MarkdownNote node provides in-graph guidance, and a SaveImage node writes results to your ComfyUI output folder.
What you provide
- Reading the workflow
Runs in ComfyUI. This workflow can't run on this page. Open it in Comfy Cloud with everything loaded, or download it for your own ComfyUI.
Try in CloudQuestions & answers
Keep phrases short (1–5 words), use uppercase for critical words, and include clear style/context cues (e.g., "on a plain white background", "centered title"). Increasing image size can help, and if the Subgraph exposes steps or guidance, modestly raise them to improve fidelity.
No. Text-to-image models can follow style hints (e.g., "bold serif", "handwritten script") but do not guarantee a specific font. For closer results, describe font characteristics and context (e.g., "retro neon sign in cursive script").
The SaveImage node writes outputs to your ComfyUI output directory. Look for the most recent image after each run. You can change naming or folder behavior by editing the SaveImage node settings.
A GPU is recommended for practical speed. CPU inference may work but can be significantly slower, especially at larger resolutions. If you encounter memory issues, try smaller canvas sizes or lower batch counts.