Text
Prompts, multimodal inputs, and split output.
Text generation creates or updates text nodes on the canvas. Use it for summaries, shot lists, prompt writing, and multimodal understanding.
Mode
Text models use Text → Text mode in the generate bar.
Inputs
Beyond the prompt in your selected node, Bragi can read upstream references:
- PDF files
- Images
- Videos
- Audio files
- Additional text or
.mdnote content via arrows
When upstream media is connected, the generate bar may show a hint such as Upstream: 1 PDF, 1 video.
If the chosen model does not support a connected media type, Run disables with a tooltip naming the conflict.
Split output
Ask the model to separate sections with this exact line on its own:
Bragi splits the response into multiple text nodes, stacked vertically. Each new node connects to the source with an edge.
This is useful for turning one planning prompt into several image prompts. See Split Text into Batch Images.
Text models
Enable models in Settings → Models → Text Models. Registered text models include:
- Claude Opus 4.7
- Claude Sonnet 4.6
- Gemini 3.5 Flash
- Gemini 3 Flash
- Gemini 3.1 Pro
- Grok 4.3
- Grok 4 Fast
- Qwen 3.6 Plus
- GPT-5.4 Pro
- GPT-5.4
Multimodal support varies by model and provider.