The secret weapon for YouTube ads.Swipe what already works for your competitors.

AdSee indexes 16M+ YouTube video ads, of which 5M+ are unlisted. 3M+ are fully AI-analyzed with transcripts and visual descriptions. The database grows by roughly 150K ads per day.

Because AdSee only indexes what's currently relevant. Our database covers YouTube ads from 2025 onward — we're a newer platform, and we built the pipeline for the way YouTube advertising works now. Older tools carry archives going back to 2022, but a creative that stopped running three years ago tells you very little about what's converting today. AdSee adds 150K+ new ads every day, so coverage of active campaigns keeps deepening — including 5M+ unlisted ads most tools never index at all.

Unlisted ads are video ads uploaded privately and served through paid placements — they never appear on the advertiser's public channel or in YouTube search. Most competitor ad spend goes to creatives you can't find by browsing their channel. AdSee holds 5M+ of them.

All of them — and you don't need to match the ad's language to find it. AdSee uses semantic (meaning-based) search: type your query in English and it will surface relevant ads in Spanish, Portuguese, German, or any other language, because the search understands what the ad is about, not just the words it uses. You can still filter by language to focus on a specific market.

AdSee tracks where each YouTube ad was shown and lets you filter by country. Coverage is deepest in North America (US, Canada), Europe (UK, Germany, France, Spain and more), Australia, and Latin America (Mexico, Brazil, Argentina), with new markets added as the database grows.

This mostly affects unlisted videos — ads that are hidden from public discovery on YouTube and exist only as ad placements. Determining where these ads were shown takes additional processing, so country data for them may appear with a delay or be temporarily unavailable. We're actively expanding this coverage.

AdSee adds 150K+ new YouTube ads every day, with AI processing running continuously — so newly launched competitor creatives typically appear within a day.

Yes. AdSee transcribes every ad in the AI-analyzed index in full, so you can search the exact words spoken — a specific hook, claim, price point, or guarantee — and get every competitor using it.

Yes. AdSee generates an AI description of the visuals of every ad in the AI-analyzed index, so a plain-language query like "dog playing on a beach" returns the ads containing that footage. No keywords or tags required.

Yes. AdSee shows which countries each ad ran in, and you can filter search results by market — useful for finding what a competitor is testing in your region versus scaling globally.

Yes. Where an ad drives to a landing page, AdSee links the two — so you can evaluate the creative and the funnel together instead of judging the video in isolation.

We estimate the media spend behind each ad from its view count and the countries it ran in, using country-level cost-per-view benchmarks. Estimates are always shown as a range (like $220K–$560K), never a single number — YouTube pricing is auction-based, so a range is the honest answer. Our methodology page explains the full approach.

Yes. From any ad, AdSee surfaces similar ads matched on spoken content, visuals, or both, with filters for industry, language, duration, and country to narrow the set.

Swipes is a one-click save: any ad you find in search goes into your Swipes, ready whenever you need it. Boards are for organised research — create a folder for each research task and a board for each competitor, then add the ads that need a closer look. Every ad you save stays available even if it is later removed from YouTube.

Open any board and ask for a report. AdSee's AI analyses every ad on it and writes a competitor creative report: the top hooks and opening lines, the core angles and messaging, the offers and calls to action, and the formats and creative patterns — with each finding linked to the ads behind it.

AdSee AI MCP is a Model Context Protocol server at https://mcp.adsee.ai/mcp that connects AI agents like Claude and ChatGPT to AdSee. Once it's connected you can ask your agent about any brand's YouTube ads in plain language: it searches the 16M+ ad database, pulls spend estimates, hooks and landing pages, reads your boards and can request creative reports. You sign in once with your AdSee account (OAuth) — there is no API key. Any agent that can add a streamable-HTTP MCP server with OAuth sign-in can use it.

Add AdSee as a custom connector in Claude: 1) Open Connectors: In any Claude chat, click + and choose Connectors. 2) Add a custom connector: Choose Add connector → Add custom connector. 3) Paste the AdSee address: Name it AdSee, paste https://mcp.adsee.ai/mcp and click Continue. 4) Sign in and approve: Claude opens an AdSee sign-in. Approve it and you’re connected. If Claude asks before it uses AdSee, choose Allow — nothing runs until you do.

Add AdSee as a connector in ChatGPT's developer mode: 1) Turn on Developer mode: Open Settings → Apps & Connectors → Advanced and switch on Developer mode. 2) Create a connector: Back in Apps & Connectors, click Create. 3) Paste the AdSee address: Name it AdSee, paste https://mcp.adsee.ai/mcp, pick OAuth and click Create. 4) Sign in and approve: Approve the AdSee sign-in, then enable AdSee from the + menu in a new chat. In developer mode ChatGPT asks before each AdSee action — choose Confirm to let it run.

Your agent only works with your AdSee research — searches, boards and reports. Billing, team and account settings always stay in the app. It can search ads, open ad details, read your boards, save ads to boards and request creative reports (reports use AI tokens, as they do in the app). The connection only exists after you approve the AdSee sign-in, and you can remove the connector in your agent at any time.