A research and targeting intelligence platform for ChatGPT Ads, built around a captured dataset of sponsored ad cards observed inside ChatGPT. It provides a plugin (via MCP) for Claude, ChatGPT, or Codex that lets an AI assistant retrieve advertiser evidence, inspect captured creatives and prompts, map nearby competitors, and draft/validate \"context hints\"—natural-language descriptions of a buyer conversation used for ChatGPT Ads targeting . The dataset is maintained by a sibling ChatGPT Ads Library that probes ChatGPT with realistic prompts and records sponsored cards . It also offers a dashboard for agencies with market views, data requests by country, and a library for broad research . Pricing (official): Single Brand $19 (one company), Agency $49 (multiple client brands), both including the full campaign intelligence workflow; free limited usage with no card . Critical caveats: The platform explicitly states its evidence is a sample, not a complete record; inferred hints are analytical reconstructions, not private Ads Manager data . It does not publish campaigns, manage bids, or access private ad account data—performance must be measured in the client\'s own Ads Manager . Independent third-party reviews were not surfaced, and the product is very new (content dated mid-to-late 2026).
Pros & Cons:
Pros: The evidence-first framing is unusually honest for an ad intelligence tool—it separates \"observed\" (captured ads and prompts), \"inferred\" (reconstructed audience/intent patterns), and \"validated\" (the user\'s own drafted hint scored against evidence), and links back to source records wherever possible . The MCP plugin workflow is genuinely agent-native: product context stays in the Claude/ChatGPT/Codex conversation, and the plugin brings in advertiser evidence and prompt data without forcing a context switch . The context-hint writing methodology is concrete—a four-part formula (buyer/use case + need + situation + material fit) with before/after examples and editing rules, grounded in OpenAI\'s guidance that hints are thematic signals rather than keywords . Agency features address real workflow needs: separate client markets, pinned competitor sets, country-specific data requests, and provenance in client briefs . Pricing is transparent ($19/$49) and the free tier allows real evaluation without a card .
Cons: The core dataset is a sample of captured ads, not a complete record of ChatGPT\'s ad inventory—so absence of evidence is not evidence of absence, and niche coverage may be uneven . Inferred hints are reconstructions, not proof of what advertisers actually submitted; the platform itself warns against treating them as exact settings . It does not measure performance—no impressions, clicks, spend, or conversions—so it cannot tell you whether a targeting direction actually works; validation must happen in the client\'s own ad account . Independent third-party reviews are absent, and the product is very new (2026), so reliability, dataset freshness, and long-term support are unverified. The plugin requires connecting a private MCP URL to an external AI client, which some teams may find adds a setup step and a data-sharing consideration . And the market is nascent—ChatGPT Ads itself is an emerging surface, so the tool\'s value depends on how widely that ad product is adopted.
Who is it for?:Agencies and in-house PPC/growth teams that are already running or planning ChatGPT Ads and want evidence-backed targeting research rather than keyword guessing; teams that use Claude, ChatGPT, or Codex for campaign planning and want ad-market evidence inside that conversation via MCP; agencies managing multiple client brands that need separate markets, country-specific evidence, and a research trail they can show clients . Not suited for teams not yet using ChatGPT Ads, anyone seeking performance measurement or bid management (the tool explicitly does neither), users requiring extensive independent third-party reviews before adoption (none surfaced), or organizations uncomfortable connecting a private MCP endpoint to an external AI assistant.