Company Enrichment API: What It Returns, and What to Use When You Need More

If you're searching "company enrichment api," you probably have a signup form with one field, a work email or a domain, and you want to know who just walked in. A company enrichment API takes that domain and hands back a record: industry, headcount band, location, maybe funding and tech stack. That's a well-defined job and a whole category of vendors does it.

This piece is for the point where that record stops being enough, typically when the thing you're building is an AI feature that has to write, recommend or explain in a way that fits this particular customer. A row of firmographics can route a lead. It can't tell your model what the customer actually sells, who they sell to, or what their refund policy says. This is a developer-focused look at the gap, and at a different kind of enrichment: indexing the customer's own website and querying it.

Quick Answer

A company enrichment API returns structured fields about a company, looked up by domain or email. Those fields are good for segmentation, lead scoring and routing. They are thin as grounding for generated output. The alternative is to treat the customer's own website as the enrichment source: crawl it, chunk and embed it into a private per-customer index, then query that index with a REST call whenever your product needs context. You get passages and cited answers in the customer's own words, rather than a fixed set of fields.

What a Company Enrichment API Actually Gives You

The category is easy to describe without naming anyone. You send a domain (or an email, from which you derive the domain), and you get back a JSON record. The usual shape:

• Firmographics. Industry, employee count band, headquarters, founding year.
• Commercial signals. Funding stage, revenue band, sometimes technologies detected on the site.
• Contacts and org data. Depending on the vendor, people, titles and email patterns.

Which fields you get, how fresh they are and what each lookup costs differ a lot between providers, so check each vendor's own docs rather than a roundup's summary. What doesn't change is the shape: a flat record of attributes, decided by the vendor, identical for every customer.

That shape is exactly right for a few jobs:

1. Lead routing and scoring. Send enterprise-sized signups to sales, self-serve the rest.
2. Form shortening. Ask for a domain, fill in the rest.
3. Segmentation and reporting. Cohort by industry and size.

Where Firmographics Stop Being Enough

Now imagine your product is a SaaS tool and you want an AI feature to draft a welcome email, suggest a first project, or write a marketing blurb for the new customer. You enrich the signup and get:

{ "industry": "Wholesale", "employees": "51-200", "country": "US" }

Feed that to a model and you get generic output with the right adjectives. The model still doesn't know:

• what the company sells, and to whom;
• the words they use for their own product (a "job" versus a "ticket" versus a "work order");
• their policies, pricing model or positioning;
• what changed on their site last month.

Those facts are not in a firmographic record because a firmographic record is built to be the same shape for every company. They are on the customer's own website, in their own words. That's the enrichment source a generation feature actually wants.

Enrichment as Retrieval: Index the Whole Site

The alternative is not a richer field list. It's changing what "enrich" means: instead of looking up attributes, build a searchable index of what the customer has published and query it.

With the white-label RAG API, that is one provisioning call at signup:

curl -X POST "https://context-link.ai/api/v1/api_accounts" \
  -H "Authorization: Bearer $PROVISIONING_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "uid": "cust-8841",
    "label": "Acme Supply Co",
    "site_url": "https://acme.com"
  }'

uid is your own customer id, which makes the call safe to retry: a replay returns the same account and token. The customer's site is crawled into chunked, embedded passages in a private index that no other customer's queries can reach. Everything is covered in the create-account docs.

Then, whenever your product needs customer context, you query with that customer's token:

curl "https://context-link.ai/api/v1/context?query=what+does+the+company+sell" \
  -H "Authorization: Bearer $CUSTOMER_TOKEN"

You get the matching passages back as markdown to drop into your own prompt. If you'd rather have a short answer to render directly, GET /question returns a grounded paragraph with citations. Both are covered in getting context for one customer.

Side by Side

	Firmographic enrichment API	Whole-site index per customer
Input	A domain or email	A site URL (plus anything you push)
Output	A fixed record of attributes	Passages and cited answers, on demand
Shaped by	The vendor's schema	Whatever the customer published
Good for	Routing, scoring, segmentation	Grounding generated copy, recommendations, onboarding
Freshness	Vendor's refresh cycle	Re-synced from the customer's site
Question it answers	"What kind of company is this?"	"What does this company say about X?"

They aren't rivals. Enrichment records are cheap structured signals for deciding what to do with a signup. A site index is grounding for deciding what to say to them. Plenty of products will want both.

Practical Notes for Building It

Don't block signup on the crawl. Crawling is deliberately slow, so the index isn't ready the moment the account is created. Provision, let the user carry on, and wait for the api_account.index_ready webhook before you lean on retrieval. The webhook docs cover the HMAC signature check and the at-least-once delivery contract, so make your handler idempotent.

Know the limits. Per-customer site crawls are capped at 500 pages. A customer with no website, or one you also want to enrich with their own documents, can have content pushed to their index instead.

Use structured output when you need citations. Adding response_format=structured to either endpoint returns the markdown plus chunk-level source metadata, which is what you need to show where a claim came from.

Keep your enrichment vendor if you have one. Nothing here replaces a firmographic lookup for scoring. This only adds the layer that lookups can't provide.

Pricing, Briefly

White-label is $50 a month per bundle of 10 customer accounts, with no token or usage fees on top. Compare that against what you'd spend assembling the same thing yourself: a crawler, chunking, an embedding pipeline, a vector database isolated per tenant and scoring that you tune yourself. If you're weighing build against buy, our RAG as a service buyer's guide covers the trade-offs, and how a RAG pipeline works shows what you'd be taking on.

Frequently Asked Questions

What is a company enrichment API?

A company enrichment API takes a domain or email address and returns structured information about the company behind it, such as industry, size and location. It's used to qualify leads and fill in forms without asking the user.

Is company data enrichment the same as retrieval?

No. Enrichment looks up a fixed set of attributes. Retrieval searches a body of text and returns the passages that match a question. When the goal is grounding an AI feature in what a specific customer says, retrieval over their own site is the better fit.

Can I enrich a customer with their website instead of a vendor record?

Yes. Provision an account with the customer's site_url, and their site is crawled into a private index you can query by REST with that customer's token.

Do I still need a firmographic provider?

Possibly. If you score leads on headcount or funding, a firmographic source still does that job. A site index complements it for generated content.

How do I know when the index is ready?

Subscribe to the api_account.index_ready webhook, or poll the account's status. The first crawl can take a while, so design your flow not to wait on it.