For developers
Waldo for developers: brand intelligence as infrastructure
200+ REST endpoints and a native MCP server, same auth and same response shapes, so you build the integration once.
The short answer
Waldo Build is a brand intelligence API plus a native MCP server: 200+ endpoints across brand activity, category trends, and audience conversation, with the same auth and response shapes both ways. Sign up at /build/get-started, let the setup wizard mint a key, and add one block to mcp.json.
The ask usually lands on engineering as “can we monitor our competitors’ ads and what people say about us.” The honest options have been grim: scrapers that break every time a platform ships a UI change, social APIs with shifting terms, or an enterprise suite whose export button is the API. You end up maintaining plumbing instead of building the thing you were asked for.
Waldo Build is that plumbing, run as a product. Tracked brands, categories, and audiences refresh on a 24-hour cycle; Discover and enrichment endpoints hit the platforms live at request time. Everything captured is stored permanently, text, images, video, and metrics, even if the original is deleted. You get it as REST or as a native MCP server with the same auth and the same response shapes, so neither surface is a second-class citizen.
The plays
Cron plus ?since= is the whole sync job
Poll /brands/{id}/ads/active?since= on a schedule and you get only what changed since your last pull. Diff, store, and fire webhooks from your own infrastructure, where you control retries and routing.
One block in mcp.json
Add the Waldo server to mcp.json and the same tools, discover_ads, brand_overview, audience_insights, category_trends, are live in Claude, ChatGPT, and Copilot. Org-wide connector install means you do it once for the whole company.
Pick your layer per call
Every call chooses raw source material, posts, ads, page snapshots, verbatims, or scored analysis. Build your own models on the raw layer at about 1 credit per 20 results, or take scored analysis at about 5 credits when the synthesis itself is what you need.
Retire the scraper stack
Instagram, TikTok, Reddit, X, YouTube, Meta, and LinkedIn, plus Meta and Google ad libraries tracked daily per brand, landing page snapshots, and GWI audience panels, all behind one auth. When a platform changes its markup, that is Waldo’s problem.
Questions teams ask
How does auth work?
Sign up at /build/get-started and the setup wizard mints an API key. Keys are scoped, so you can cut separate keys per service or per team, all drawing from one org-level credit pool. The same credentials work for REST and MCP.
Are the REST and MCP responses actually the same?
Yes. The MCP server is native, not a wrapper bolted on later: same auth, same response shapes. A tool call to discover_ads returns what the corresponding REST call returns, so you can prototype in chat and move to code without remapping fields.
What does usage cost at volume?
Pricing is credit-metered: roughly 1 credit per 20 results on the raw layer and about 5 credits for analysis calls, with volume discounts as usage grows. There are no per-seat fees or per-brand minimums; the only fixed cost is an annual activation fee for entities Waldo isn’t already tracking.
Keep reading
Put Waldo behind your agents
Brand, category, and audience intelligence over 200+ API and MCP endpoints. Sign up, mint a key, and run it against the brands you actually track.