Can My Agent Read Inline Images from a Google Doc?

Last updated: April 14, 2026

Yes — Gumloop agents can read inline images from Google Docs. Add the Google Docs tool to your agent, and the read_doc tool will return both text content and inline image URIs, which the agent can then process with its multimodal LLM.

What Works and What Doesn't

How images are handled depends on whether you use an agent or a workflow:

Approach

Text

Inline Images

Agent with Google Docs tool

✅ Full text with formatting

✅ Returns image URIs via inlineObjects

Workflow with Google Docs Reader node

✅ Full text with formatting

❌ Text only — images are skipped

If your Google Doc contains screenshots, diagrams, or any embedded images that you need an LLM to analyze, use an agent.

How to Set It Up

  1. Go to your agents and open (or create) the agent you want to use.

  2. Add the Google Docs integration as a tool. If you haven't connected it before, see How to Give Your Agent Access to an Integration.

  3. In the chat, give your agent the Google Doc URL and tell it what you need — for example: "Read this Google Doc and summarize all the content, including any screenshots or images."

  4. The agent will call the read_doc tool, which returns the document's text along with any inlineObjects (inline images) and positionedObjects (floating images). The agent's multimodal model can then interpret the images and include them in its response.

Example prompts you can try:

  • "Read this Google Doc and summarize the text and all embedded screenshots: [URL]"

  • "Parse this doc and extract any data from the inline images: [URL]"

  • "Summarize this doc — there are Slack conversation screenshots inline that I need you to read: [URL]"

Common Pitfall: Adding the Integration

If your agent replies that it can't access Google Docs, the most common cause is that the Google Docs tool hasn't been added to the agent's tool list. Updating the agent's instructions alone is not enough — you must explicitly add the integration from the agent's Tools panel.

See How to Give Your Agent Access to an Integration for step-by-step instructions.

What About Notion Docs?

The same agent-based approach works for Notion pages. Add the Notion tool to your agent, and it can read Notion page content in a similar way. If you run into permission issues with Notion, see Gumloop Can't Find or Read My Notion Page or Database.

Still Need Help?

If this didn't resolve your issue, reach out to support at support@gumloop.com.

Related Docs