Skip to main content
Cookbooks are battle-tested recipes for the research tasks people run against GOI most often — supplier shortlists, lookalikes, market sizing, niche and deep-tech discovery, and agent workflows. Each one is a fixed pattern: you supply one or two inputs, everything else stays locked in, and you get the same clean result every time.

​
Query templates

The GOI API is super powerful. A single POST /v2/search endpoint can run a semantic, full-text (BM25), hybrid, and similarity search across ~20M web-verified organizations, with structured filters for geography, sector, size, and more. That flexibility is the point but it also means there are many ways to phrase any given query, and the gap between an okay result and an excellent one lives entirely in the parameters. Query templates are the recommended way to harness that power. Each one distills a common use case into a fixed, validated parameter recipe: the disciplined combination of describe, keywords, filters, and more that reliably returns the right organizations. Think of them as the canonical “best way to do X with the GOI API”: copy the recipe, swap in your inputs, and ship; no need to rediscover the winning combination yourself. Every template includes a copy-paste Python example against POST /v2/search or POST /v2/stats.

Supplier discovery

Given a product + location, return a supplier shortlist (makers + distributors).

Find lookalikes

Given a reference organization, find similar organizations, competitors, targets, prospects.

Market sizing

Count and break down a sector across geographies, exact, reproducible numbers.

Niche & deep-tech finder

Find the organizations that actually do one specific thing — a niche product or frontier tech.

​
MCP & agents

Recipes for driving GOI through the MCP connector in Claude, Cursor, or any agent.

Research workflows

Chain tools end to end — resolve a place, size a market, pull the players, profile one.

Agent skill

A drop-in skill so an agent applies the right parameters automatically.
Reproducible within an index version. Same inputs + same template + same GOI snapshot → same results. Results change as the corpus refreshes; to freeze output, snapshot the returned domains and re-hydrate later via get_organization_details.