Google Images — Localised Image Search That Saves Straight Into Your Session

May 16, 2026

Ask for images the way you'd ask a researcher — by topic, license, colour, recency, or aspect ratio — and Alfrada pulls the matches from Google, downloads the top five into your session, and indexes them so the Presentation tool, Video Editor, and Report Generator can drop them in with one more sentence. Results auto-localise to your country, so a search from Berlin returns the German index by default. No Pexels-style stock-only constraint — this is the open web.

What you can do

• Search Google Images by query — google_images runs the search via SearchAPI and returns ranked results with thumbnail, source page, dimensions, and a credit line.
• Top results auto-save to your session. Default 5 images download into out/ as real files (set download_count to a different number, or 0 for metadata-only). Each saved image gets indexed alongside its title, source name, source link, and a ready-to-render credit line.
• Auto-localised results. Country and language flow from your detected location — gl=de and hl=de for Berlin, gl=gb and hl=en for London, and so on. Override with the location field when you want a different regional bias (e.g. "Tokyo, Japan").
• Filter by image size — 16 size options including large, medium, icon, fixed-resolution thresholds (larger_than_400x300 up through larger_than_1024x768), and megapixel thresholds (larger_than_2mp up through larger_than_70mp).
• Filter by colour — 15 colours including color, black_and_white, transparent, plus 12 named hues (red, orange, yellow, green, teal, blue, purple, pink, white, gray, black, brown).
• Filter by image type — photo, clipart, line_drawing, gif, face.
• Filter by aspect ratio — square, tall, wide, panoramic.
• Filter by recency — last_hour, last_day, last_week, last_month, last_year. Useful for breaking-news visuals or fresh product shots.
• Filter by usage rights — creative_commons_licenses (CC-licensed only) or commercial_or_other_licenses (commercial-use friendly). Default returns everything Google indexes.
• SafeSearch control — active (strict), blur (default), or off.
• Pagination — page flips through pages of results without re-running the search.
• Custom filename prefix — pass output_filename_prefix to control how saved files are named, or let the tool derive a prefix from the query.
• Inline rendering with credits. Each saved asset returns an out/[filename] path, alt text, title, source name, source link, and credit line, ready to drop into markdown — ![alt](out/filename) for the image, then *title* — [source name](source link) (via Google Images) for attribution.

Where this shows up

• Building a slide deck and your pitch needs three real photos of the Brandenburg Gate at night. "Find five wide-aspect, large-size night shots of the Brandenburg Gate. Save them and drop them into the deck." The presentation tool slots them in with attribution.
• You want a credit-clean blog hero. "Find a panoramic landscape photo of the Lofoten Islands — Creative Commons only, save the top three, list the photographers." Filenames, source links, and credit lines come back with the files.
• Quick visual research before a pitch meeting. "Pull thumbnails of the latest Tesla Cybertruck colour options — last month only, metadata only, no downloads." Set download_count=0 and you get a ranked list of links without filling your session.
• Fresh news visuals. "Find images of the SpaceX Starship launch from this week — large size, photo type." Recency + image_type narrow to current photography, not stock.

Try it

• "Find five large images of art deco interiors, save them to my session, and credit the photographers."
• "Pull 10 Creative Commons photos of the Tokyo skyline at sunset, wide aspect, and list the source URLs."
• "Search for transparent PNG icons of a cardboard box — image_type clipart, color transparent."
• "Find 3 line drawings of the human respiratory system for my biology slide."
• "Get me a panoramic photo of the Lofoten Islands, last month, save the top one."
• "Find 8 black and white photos of vintage cameras — square aspect ratio."

Heads up

• Downloads are capped. Each image gets up to 15 seconds to download and 12 MB before the tool gives up and moves on. Hotlink-blocked or oversized images get skipped — the rest still save.
• Embedding paths matter. When you reference saved images in chat, use the out/<filename> relative path the tool returns. Plain source-page hyperlinks won't render the image inline.
• cr (country restrict) is intentionally omitted. SearchAPI's gl + hl already route the search to the local index. Adding cr over-filters image queries and starves the result set — so the tool skips it.
• Attribution travels with every file. Each saved asset comes with a credit line; render it visibly when publishing. Google Images surfaces images from across the web — most images are not Creative Commons by default. Use usage_rights=creative_commons_licenses when you need redistribution-safe sources, or check the source link before publishing.
• Pairs with the rest of the toolchain. Saved images are immediately usable by presentation (slide images), video_editor (render_slide for slideshow MP4s), and report_generator (embedded report images) in the same session.