> ## Documentation Index
> Fetch the complete documentation index at: https://docs.shanone.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Google Gemini

> Deep research and image generation via the Gemini API

# Google Gemini

Shanone's Gemini integration gives your agent 6 tools for two specialized capabilities the chat model itself doesn't cover: long-running multi-step research and image generation.

## Getting Started

<Steps>
  <Step title="Get a Gemini API key">
    Create an API key from [Google AI Studio](https://aistudio.google.com/apikey).
  </Step>

  <Step title="Add the key in Shanone">
    Open **Integrations** in the Shanone dashboard, select Google Gemini, and paste the API key.
  </Step>

  <Step title="Retry your request">
    Once saved, `shanone_execute_tool` calls for `gemini_*` tools will succeed.
  </Step>
</Steps>

## Available Tools

Shanone provides **6 tools** for Google Gemini, organized into these categories:

<AccordionGroup>
  <Accordion title="Deep research">
    `gemini_deep_research`, `gemini_get_research_status`
  </Accordion>

  <Accordion title="Image generation">
    `gemini_generate_image`
  </Accordion>
</AccordionGroup>

## Common Use Cases

<CardGroup cols={2}>
  <Card title="In-depth topic research" icon="telescope">
    Kick off `gemini_deep_research` on a complex question and get a synthesized, cited report back
  </Card>

  <Card title="Concept art & illustrations" icon="image">
    Generate an image from a text description for a mockup, blog post, or presentation
  </Card>

  <Card title="Background research jobs" icon="clock">
    Start a research task and poll `gemini_get_research_status` while the agent continues other work
  </Card>

  <Card title="Visual asset drafts" icon="palette">
    Quickly draft several visual concepts to compare before commissioning final artwork
  </Card>
</CardGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="gemini_deep_research doesn't return a final answer immediately">
    Deep research runs asynchronously for complex queries — poll `gemini_get_research_status` with the returned research ID until it reports completion.
  </Accordion>

  <Accordion title="gemini_generate_image produces text or details that don't match the prompt">
    Be explicit about style, composition, and content in the prompt — vague prompts tend to produce generic results; iterate with a refined prompt rather than assuming the first generation is final.
  </Accordion>

  <Accordion title="Tool calls fail with an invalid API key error">
    Gemini API keys from Google AI Studio are separate from a Google Cloud service account — confirm the key was generated at aistudio.google.com and reconnect if it was revoked or regenerated.
  </Accordion>
</AccordionGroup>