Overview
To use this template, ensure you have a Gemini API key. You need to provide a file (image, audio, or PDF) and a prompt. Optionally, specify a model and an output folder for saving the response.
Variables
|
Name |
Reference |
Type |
Required |
Default |
Options |
Description |
|---|---|---|---|---|---|---|
|
Model |
GEMINI_MODEL |
Select |
➖ |
gemini-2.5-flash |
Gemini 2.5 Flash: gemini-2.5-flash
|
Optional model to use. Default is gemini-2.5-flash. |
|
Prompt |
GEMINI_PROMPT |
Alphanumeric |
✅ |
- |
- |
Provide the prompt to send to the API along with the file. |
|
Match Type |
GEMINI_MATCH_TYPE |
Select |
➖ |
exact_match |
Exact Match: exact_match
|
|
|
Source File Name |
GEMINI_SOURCE_FILE_NAME |
Alphanumeric |
✅ |
- |
- |
|
|
Source Folder Name |
GEMINI_SOURCE_FOLDER_NAME |
Alphanumeric |
➖ |
- |
- |
If the file that Gemini is in a subfolder |
|
Destination File Name |
GEMINI_DESTINATION_FILE_NAME |
Alphanumeric |
✅ |
- |
- |
Name to save the response file locally. |
|
Destination Folder Name |
GEMINI_DESTINATION_FOLDER_NAME |
Alphanumeric |
➖ |
- |
- |
Optional folder path to save the response file. |
|
Enable Google Search |
GEMINI_ENABLE_GOOGLE_SEARCH |
Boolean |
➖ |
FALSE |
- |
Connects the Gemini model to real-time web content |
|
Enable Code Execution |
GEMINI_ENABLE_CODE_EXECUTION |
Boolean |
➖ |
FALSE |
- |
When enabled, the model executes Python code and returns the results. Disable this option if you want the model to include the code itself in the response. |
YAML
Below is the YAML template
source:
template: Gemini - Generate Response From File
inputs:
GEMINI_MODEL: gemini-2.5-flash
GEMINI_PROMPT:
GEMINI_MATCH_TYPE: exact_match
GEMINI_SOURCE_FILE_NAME:
GEMINI_SOURCE_FOLDER_NAME:
GEMINI_DESTINATION_FILE_NAME:
GEMINI_DESTINATION_FOLDER_NAME:
GEMINI_ENABLE_GOOGLE_SEARCH: 'FALSE'
GEMINI_ENABLE_CODE_EXECUTION: 'FALSE'
type: TEMPLATE
guardrails:
retry_count: 1
retry_wait: 0h0m0s
runtime_cutoff: 1h0m0s
exclude_exit_code_ranges:
- 10
- 11
- 14
- 205
- 204