Gemini Spark can now control Google Photos: Features and Access

Lisa Ernst · 05.09.2026 · Artificial Intelligence · 9 min.

Google is enhancing Gemini Spark with a significantly deeper connection to Google Photos. Instead of just searching for images, the AI agent can now perform multi-step photo workflows: find and curate shots, edit images as new copies, create albums, prepare shares, extract text from photos, and schedule recurring tasks.

The crucial catch at launch: The Gemini Spark Google Photos integration is rolling out gradually and, according to Google, is currently only available to eligible users aged 18 and over in the USA. The Google Photos features in Spark will initially only work in English and on the mobile Gemini app and web. For users in Germany or Switzerland, as of September 5, 2026, there is no regular access yet.

Source: scikit-image.org / Lav Varshney (CC0)

In Brief

• More than search: Spark can combine Google Photos tasks in a multi-step workflow.
• Edit without overwriting: Changes are saved as a new copy; the original remains untouched, according to Google.
• Albums and shares: Spark can create new private or shared albums and forward links via connected apps. Confirmation is required for sensitive sharing actions.
• Automation: Recurring tasks such as monthly highlight albums or sorting receipt photos can be scheduled.
• Limited launch: Google Photos in Spark is currently US and English exclusive and requires Spark eligibility and a connected Google Photos library.

What's new with the Gemini Spark Google Photos integration?

Google Photos could already be connected to Gemini before Spark to find specific photos, videos, and albums based on people, places, dates, or image content. However, the normal gallery connection has clear limitations: Google's help page explicitly states that Gemini apps within the gallery app cannot use editing functions and cannot create new albums or modify existing ones.

This is precisely where Spark comes in. Google describes Spark as a personal AI agent for complex workflows. With the new Google Photos connection, a single request can trigger multiple steps. An inquiry like 'Find the best vacation pictures, improve the beach photos, create an album, and prepare a message with the link' becomes a cohesive task instead of multiple manual individual steps.

What tasks can Spark perform in Google Photos?

1. Search, sort, and curate photos

Spark can search for shots by subject, location, date, or event. Additionally, the agent can select the best images and filter out duplicates. This is particularly interesting for very large libraries where a simple search might yield results, but the actual selection still takes time.

Source: scikit-image.org / Rachel Michetti (CC0)

Symbolic image for subject-based search: Spark can search Google Photos content by specific subjects and then further process the matching shots.

2. Enhance images and create new versions

Google mentions quick quality improvements, corrections, stylized variations, and the creation of collages and videos as supported actions. The saving logic is important: Spark does not overwrite the original image. For an edit, a new copy is created and saved in Google Photos. Therefore, those who want to review results can compare the original and the AI-edited version separately.

For classic Gemini photo editing without Spark, there is already a dedicated guide on Zerlo Guide to photo editing with Google Gemini. Spark primarily differs from this in that editing can be combined with organization, sharing, and other connected apps within an agent workflow.

Source: NASA / scikit-image.org (Public Domain)

People and recurring faces are typical search criteria in large photo libraries. The image serves only as a real symbolic example and does not show a Gemini interface.

3. Create albums and prepare shares

Spark can place selected shots into a new private or shared album. Newly created albums are initially private, according to Google. If an album is to be shared or, for example, a link is to be sent via email, Spark requires confirmation for certain actions. This reduces the risk of the agent sending personal photos externally without final approval.

4. Transfer information from images to other apps

The integration is not limited to image management. Spark can extract text from images and use the information further, for example, to summarize content or transfer it to a Google Document. Google also provides an example where a photo of a local concert schedule is matched with Google Calendar to find a suitable appointment. This turns the photo library into a source of information for other connected apps.

Source: SpaceX / scikit-image.org (Public Domain)

Event, location, and time are also search criteria. Spark can find matching shots and then include them in a larger workflow.

5. Automate recurring photo tasks

The biggest difference from a one-time chat request is time control. Spark can take over recurring tasks, for example, selecting the best food photos every weekend, creating a collage from them, or creating a highlight album monthly. Google also mentions sorting receipt photos as a possible automation scenario.

Such tasks should not run blindly. Google generally points out with Spark that the agent is experimental and can make mistakes. Manual control is advisable, especially with sensitive data or actions that will be shared later.

Source: NASA / scikit-image.org (Public Domain)

Large image collections can contain very diverse subjects. Spark's strength lies in combining search, selection, and follow-up actions in a single task.

Prerequisites: Who gets access?

General Spark availability and the new Google Photos feature are two different levels. Even if Spark is generally available for an account, Google Photos may still be missing because the integration is activated separately and gradually.

Germany and Switzerland: The new Google Photos feature in Spark is expressly limited to users in the USA. A change of language alone is therefore not sufficient. Furthermore, Google currently excludes Switzerland and the European Economic Area from general Spark availability. Google has not yet announced an official date for the international launch of the Google Photos integration.

How to enable Google Photos in Gemini Spark

If your account meets all requirements and the feature has already been rolled out to you, setup involves two steps: first connect Google Photos with Gemini, then open Spark.

1. Log in with the same personal Google account: Open the Gemini mobile app or gemini.google.com and check if you are logged in with the account you also use for Google Photos.
2. Initiate Google Photos: Ask Gemini to find a photo from your library. Google recommends using <strong>@Google Photos</strong> in your prompt to clearly address the connected app.
3. Confirm connection: If Google Photos is not yet connected and your account is eligible, Gemini will display an option to connect. Follow the on-screen instructions.
4. Open Spark: In the Gemini app, you can access Spark via the menu. On the web, switching to Spark is in the sidebar.
5. Describe workflow: Formulate not just the search, but the desired end result, such as selection, editing, album, and subsequent sharing as a coherent task.
6. Check confirmations: For actions such as creating a shared album or sending an email, Spark may ask for explicit confirmation before execution.

If Google Photos does not appear despite meeting the requirements, this does not necessarily indicate an error. Google expressly states that the rollout is gradual. An eligible account may therefore receive the feature later than another.

Privacy: What happens to your photos?

With this integration, two layers of privacy work together: Google Photos and Gemini Apps. Google explains for the Gemini functions within Google Photos that personal data from Photos will not be used for advertising and that generative AI models outside of Google Photos will not be directly trained with the personal image and video data from the Photos library.

However, if Google Photos is connected to another Google service such as Gemini Apps, its own policies apply to the processing in that service. Google explicitly points out that with connected services, for example, summaries, conclusions, or generated media based on content from the Photos library, can be used for training. In addition, <strong>Keep Activity</strong> is a prerequisite for Spark. With the setting enabled, Google stores Gemini activity and, according to the Gemini Apps Privacy Hub, can use it to improve its services, including training generative AI models; human reviewers may also be involved.

Therefore, restraint is advisable for particularly private content. Spark should not be understood as invisible local automation, but as a connected cloud agent. Google itself advises against entering sensitive information directly into Spark tasks and recommends user control for sensitive or unintentional actions.

Gemini Spark, Ask Photos, and normal Gemini search: the difference

Google now offers several AI layers around photos that are easily confused. <strong>Ask Photos</strong> is a feature within Google Photos for natural questions about your own library. The normal connection of Google Photos with Gemini Apps allows you to find and discuss photos in Gemini. Spark goes a step further and executes multiple actions as an agent workflow.

You can also find a broader overview of already available functions in the Zerlo article on AI Features in Google Photos. For Spark, however, the current Google documentation is authoritative, as availability and scope of functions are still heavily dependent on the account and rollout.

The main limitations at launch

• No global launch: The Google Photos integration in Spark is currently limited to the USA.
• English only: The new Photos features in Spark are initially only available in English.
• Paid access: In the USA, Spark requires Google AI Pro or Ultra.
• Personal accounts: Work and school accounts are currently excluded from Spark.
• Keep Activity is mandatory: Anyone who does not wish to enable this activity setting cannot use Spark at this time.
• Gradual rollout: Even a fundamentally eligible account may not yet see Google Photos in Spark.
• Agents can make mistakes: Google describes Spark as a learning system and recommends particular caution with sensitive tasks.

FAQ

Conclusion

The <strong>Gemini Spark Google Photos Integration</strong> turns the previous photo search into a true agent workflow for the first time: Spark can find, select, edit images as copies, organize them into albums, reuse information from them, and schedule recurring tasks. The practical added value lies less in a single new photo function and more in the combination of multiple steps in one task.

For most German-speaking users, however, this is currently still a pipe dream. The launch is limited to the USA and English, requires an eligible Pro or Ultra account, and necessitates Keep Activity. Those who gain access should consciously check permissions and sensitive tasks and not let the agent work with private content unsupervised. Google has not yet announced a date for a broader international rollout.