Proxy vs Foreplay: live ad research or a saved swipe file

Proxy vs Foreplay comes down to one question: do you need a saved swipe file, or a fast answer about what competitors are running right now? Foreplay is the better swipe file; Proxy is the faster, cheaper way to ask an AI assistant about live competitor ads.

(Proxy publishes this blog, so weigh that. Foreplay details come from Foreplay's own FAQ pages and third-party pricing breakdowns, linked inline and checked October 2026. Prices change, so confirm before you buy.)

How do Proxy and Foreplay compare side by side?

Proxy and Foreplay differ most on where the data comes from: Proxy reads the public ad libraries live, while Foreplay searches its own stored database of saved ads. Everything else in the table follows from that split.

	Proxy	Foreplay
Data source	Live public ad libraries, fetched when you ask	Foreplay's stored database, "over 200M ads" according to Foreplay's FAQ, plus ads your team saves
Platforms	Meta (Facebook and Instagram), Google (Search, Display, YouTube via the Ads Transparency Center), LinkedIn, TikTok (EU ads only)	Facebook, Instagram, TikTok, YouTube, LinkedIn "and more" per the same FAQ. We found no Google Search ad coverage
MCP / AI assistant access	Yes, it's the whole product. Every plan, including Free	Yes, via Foreplay's MCP. Uses your paid plan's API credits, per Foreplay (FAQ)
Works in	ChatGPT, Claude, Claude Code, Cursor, Codex, any MCP client	Claude, ChatGPT, other MCP clients (FAQ)
Saved boards and swipe file	No. Results live in your chat history	Yes, the core of the product
Video transcripts	Yes, on demand for any video ad you pull (1 credit each)	Yes, per Superscale's review: automated hook and ad transcription on saved ads (review)
Competitor tracking	Re-run the same ask weekly, or schedule it in your assistant	Spyder, automated brand tracking
Price	Pro $50/mo (100 credits/day), Max $100/mo (250 credits/day)	About $59, $175, and $459/mo monthly; about $49, $149, and $389/mo billed annually, according to Creatify (breakdown)
Free tier	20 credits a day, no card	No free plan. A 7-day trial that asks for a card, per Foreplay (FAQ)

What's the real difference between live and stored ad data?

Live data tells you what a competitor is running today, and stored data tells you what someone saved at some point. Both are useful; they answer different questions.

When you ask Proxy "What is Notion running on Google right now?", it goes to the Google Ads Transparency Center at that moment and brings the ads back into the chat. Nothing is cached ahead of time, so a brand nobody has ever researched works just as well as a famous one. The flip side is that Proxy can only show what the library shows. If an advertiser pauses an ad tomorrow, Proxy won't have kept a copy.

Foreplay works the other way round. According to Foreplay, ads get into its platform when users save them and when its systems collect them (how ads get in), and Foreplay's FAQ reported that the copy you save stays in your account (FAQ). That's exactly what a creative team wants for a swipe file. It's less useful when the question is "what did this niche B2B competitor launch on LinkedIn this week?", because a database is only as fresh as the last time something was saved into it.

Which tool covers more platforms?

Proxy covers the four public ad libraries marketers ask about most: Meta, Google, LinkedIn, and TikTok (EU ads, searched by exact advertiser name). Foreplay covers Meta, TikTok, YouTube, and LinkedIn in its database, but we couldn't find Google Search or Display ads anywhere in Foreplay's documentation.

That gap matters more than it sounds. Data from Proxy shows that across the roughly 179,000 ads Proxy returned to marketers' AI assistants between February and October 2026, 16.5% came from Google and 3% from LinkedIn. For a search marketer or a B2B team, Google text ads are often the clearest read on a competitor's positioning, because the headline has to say the offer in 30 characters. Our guide on how to check competitors' Google ads shows what that looks like.

Where Foreplay pulls ahead is depth on social video. If your job is TikTok and Reels creative, its stored, filterable library (filters for niche, format, language and running duration, per Foreplay's FAQ) is built for browsing at scale. Proxy's TikTok coverage is narrower: EU-delivered ads only, and you need the advertiser's registered name, not a keyword.

How do the two MCP servers differ?

Both Proxy and Foreplay now plug into ChatGPT and Claude through MCP, but Proxy's MCP is the product while Foreplay's is a window into a dashboard you already pay for.

Foreplay's MCP

According to Foreplay, its MCP lets Claude or ChatGPT search ads, pull competitor intel, browse your swipe file, and check your tracked brands without leaving the conversation (how it connects). It draws on your plan's API credits rather than charging extra. Analysis from Creatify puts that at 10,000 API credits a month on monthly plans and 20,000 on annual (source). The big win: the AI can read your boards, so "summarise the hooks in our Q4 skincare board" works.

Proxy's MCP

Proxy is a hosted MCP server at https://mcp.useproxy.dev/mcp; you sign in with OAuth and there's no API key or dashboard. One credit buys one brand lookup, one ad fetch, or one video transcript, and the free tier gives you 20 a day. A typical ask looks like this:

"Find Linear's competitors, then pull their active LinkedIn and Google ads and tell me which angle each one is leaning on."

Proxy finds the competitors from a description, fetches the live ads, and your assistant writes the comparison. Data from Proxy shows users have run over 16,000 research requests like this since February 2026, and about a third of the Meta ads they pull come from a topic search rather than a named competitor. That's a habit a stored swipe file doesn't really support.

When does Foreplay win?

Foreplay wins whenever the ads need to be kept, organised, and shared with a creative team. Be honest about which camp you're in before you switch anything.

• Boards and swipe files. Saving an ad with one click, tagging it, and dropping it into a board the whole team sees is Foreplay's core, and Proxy has nothing like it.
• Team briefs. Foreplay's higher plans turn saved ads into creative briefs. Proxy can draft a brief in your chat, but it won't live anywhere your designers work.
• Historical archive. Saved ads stay after the advertiser turns them off. Proxy reads what the libraries show today, so it can't recover an ad that's gone.
• Scheduled brand tracking. Spyder watches brands for you. With Proxy you re-run the ask or set up a scheduled task in your assistant, which works but is less hands-off.
• Seats. Workflow and Agency plans cover multiple users (up to 5 and 10 users, according to Creatify). Proxy plans are per account.

If three or more people save ads every day, Foreplay is probably worth its price on boards alone.

When does Proxy win?

Proxy wins when the job is answering a competitor question quickly, across more than one platform, without opening another tool.

• Live, cross-platform pulls. One ask covers Meta, Google, LinkedIn, and TikTok (EU), including the Google Search ads Foreplay doesn't list.
• It lives in ChatGPT or Claude. No dashboard to learn and no export step; the analysis happens where you already write. Data from Proxy found that Claude is the most common client, with 108 Proxy users connecting through Claude and 75 through Claude Code between February and October 2026.
• A real free tier. 20 credits a day with no card covers a few brands a day indefinitely. Foreplay's trial ends after 7 days and needs a card up front.
• Transcripts on any ad, not just saved ones. Ask for a transcript of a competitor's video ad you've never seen before and get the script and hook back in the chat. Proxy users have transcribed over 1,000 competitor video ads this way.
• Lower entry price for paid use. Pro is $50/mo against Foreplay's $59/mo Basic, though the two meter usage very differently.

Proxy has real limits too. There are no saved boards, TikTok is EU-only by exact advertiser name, credits reset daily with no rollover, and no tool (Proxy included) can show spend or ROAS for commercial ads, because the public libraries don't publish it.

Should you use Proxy and Foreplay together?

Yes, if you have both a research job and a creative job, using both is often the cheapest setup that covers everything. They overlap less than the "vs" framing suggests.

A pattern that works well for agencies:

1. Monday research in Claude or ChatGPT with Proxy. "What did our three main competitors launch on Meta and Google this week, and which ads have been running longest?" Long-running ads are the best free signal of what's working (see how to monitor competitor ads).
2. Save the keepers in Foreplay. The few ads worth studying go into a board so the creative team can brief from them.
3. Ask Foreplay's MCP about the board later. Once the archive builds up, your assistant can read it for patterns.

If you only need one, use the 30-second test: do you mostly save ads, or mostly ask questions about them? Savers want Foreplay. Askers want Proxy. If you're still weighing other tools, our Foreplay alternatives roundup and best ad library MCP servers comparison cover the wider field.

How do you try Proxy next to Foreplay?

You can try Proxy alongside Foreplay in about two minutes without cancelling anything. Paste https://mcp.useproxy.dev/mcp as a custom connector in ChatGPT or Claude, sign in, and ask about a competitor you already track in Foreplay. Compare what comes back with your board. Setup for every client is on the install page, and plan details are on pricing.