ComfyUI Node
Expanded Gemini Text/Image
A ComfyUI node in API/Gemini with 25 inputs and 3 outputs.
Expanded Gemini Text/Image
- config
- input_image
- input_image_2
- text
- image
- final_actual_seed
◄prompt—►
◄system_instructionYou are a helpful AI assistant.►
◄modelgemini-2.5-flash►
◄temperature1.00►
◄top_p0.95►
◄top_k40►
◄max_output_tokens8192►
◄include_imagesfalse►
◄aspect_ratioNone►
◄bypass_modeNone►
◄thinking_budget0►
◄use_proxyfalse►
◄proxy_host127.0.0.1►
◄proxy_port7890►
◄use_seedtrue►
◄seed0►
◄timeout30►
◄include_thoughtsfalse►
◄thinking_levelNone►
◄media_resolutionunspecified►
◄max_retries3►
◄retry_pattern►
CategoryAPI/Gemini
Inputs (25)
| Name | Type | Default | Description |
|---|---|---|---|
| config | GEMINI_CONFIG | — | |
| prompt | STRING | — | |
| system_instruction | STRING | You are a helpful AI assistant. | — |
| model | COMBO | gemini-2.5-flash | 15 options: gemini-2.0-flash, gemini-2.0-flash-lite, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite, gemini-3-pro-preview, +9 |
| temperature | FLOAT | 1.000–1 | — |
| top_p | FLOAT | 0.950–1 | — |
| top_k | INT | 401–100 | — |
| max_output_tokens | INT | 81921–65536 | — |
| include_images | BOOLEAN | false | — |
| aspect_ratio | COMBO | None | 11 options: None, 1:1, 9:16, 16:9, 3:4, 4:3, +5 |
| bypass_mode | COMBO | None | 4 options: None, system_instruction, prompt, both |
| thinking_budget | INT | 0-1–24576 | 0 disables thinking mode, -1 will activate it as default dynamic thinking and anything above 0 sets specific budget |
| use_proxy | BOOLEAN | false | — |
| proxy_host | STRING | 127.0.0.1 | — |
| proxy_port | INT | 78901–65535 | — |
| use_seed | BOOLEAN | true | — |
| seed | INT | 00–2147483647 | — |
| timeout | INT | 3015–300 | — |
| include_thoughts | BOOLEAN | false | — |
| thinking_level | COMBO | None | Does not work at the same time as 'thinking_budget'. if this is set, then thinking budget is ignored. |
| media_resolution | COMBO | unspecified | Set input media resolution for image, video and pdf. This changes tokens consumed. |
| max_retries | INT | 30–10 | Maximum number of retry attempts when pattern matches. 0 disables retry. |
| input_imageopt | IMAGE | — | |
| input_image_2opt | IMAGE | — | |
| retry_patternopt | STRING | Regex pattern to match in response text. If matched, retry with new seed. Leave empty to disable. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| text | STRING | — |
| image | IMAGE | — |
| final_actual_seed | INT | — |