Skip to main content
This guide is for engineers whose code or AI agents call Explorium’s AgentSource API, v1 or v2, or its MCP server. It maps each Explorium endpoint and MCP tool to CompanyEnrich, lists the behavior that changes in your code, and shows how to check the result on records you already have.
Explorium says AgentSource will be discontinued on December 2, 2026. Read Explorium’s notice for what happens to your credits and account.
Explorium behavior on these pages was checked against Explorium’s public developer documentation for AgentSource v1 and v2 on November 16, 2026. CompanyEnrich behavior comes from the V2 API reference. The full mapping is also available as JSON at /migrations/explorium.json.

Field mapping

Every Explorium business and prospect field, with its CompanyEnrich counterpart.

Search filters

Translate fetch filters for businesses and prospects.

With a coding agent

Instructions a coding agent follows to migrate your codebase.

Check the fit first

Switch company enrichment in five minutes

This is the shortest path: a business match and firmographics call, replaced by one enrichment call with the Python SDK.
1

Get an API key

Create a key at app.companyenrich.com/access-keys and export it as COMPANYENRICH_API_KEY. Check it with GET /v2/me, which is free:
2

Install the SDK

The Python SDK needs Python 3.10 or later. There are also SDKs for JavaScript, Ruby, and PHP.
3

Replace the calls

4

Update the fields you read

The field mapping covers every other field.
5

Handle no match

Explorium’s match returns 200 with a business_id of null. CompanyEnrich returns 404, and the SDK raises NotFoundError. It is not charged.

What takes longer

Migrate with a coding agent

A coding agent can do the inventory, the rewrite, and the tests. Give it this prompt in your repository:
See Migrate with a coding agent for what the agent does and the rules it follows.

Move your MCP setup

If your agent uses the Explorium MCP server, point it at the CompanyEnrich server instead. The hosted server is at https://mcp.companyenrich.com/mcp. For example, in Claude Code:
The MCP server guide covers Claude Desktop, Cursor, VS Code, and other clients, and lists the credits each tool uses. Then update any prompts or code that name Explorium tools: CompanyEnrich tools take plain values, such as a domain or a job title, so your agent does not need to look up IDs or autocomplete values first.

Endpoint mapping

Both API versions are listed. A v1 path such as /v1/businesses/match maps the same way as its v2 path. Where an Explorium call takes business_ids or prospect_ids, send the company’s domain, or the CompanyEnrich person ID, instead. Explorium IDs cannot be converted.

Company

People

Supporting APIs

Request and response differences

Authentication

From match and enrich to one call

Explorium works in two steps: match a business or prospect to get its ID, then pass the ID to each enrichment you need. v2 can also match and enrich in one call with businesses_to_match. CompanyEnrich returns the record from the lookup itself:

Response shape

Status codes

Two differences need code changes. Explorium reports a missed match inside a 200 response, as a null ID, while CompanyEnrich returns 404. Explorium returns 403 when your credits run out, while CompanyEnrich returns 402. The SDK raises a typed exception for each non-2xx status, such as NotFoundError, PaymentRequiredError, and RateLimitError.

Billing behavior that affects your code

These differences change how you write retries, guards, and tests. Prices for each endpoint are in Credits.
  • No match is free. Enrichment that finds nothing returns 404 and is not charged.
  • An empty search is charged. A search with no results costs one result’s base price, such as 1 credit for a filter search. Check a new filter with the free count endpoint first.
  • Extra company fields are expansions. The base record excludes financial, keywords, naicsCodes, technologies, and workforce. Each one you request adds 1 credit per company returned, so request only the ones your code reads.
  • Work emails are a separate call. GET /v2/people/{personId}/email costs 10 credits only when it finds an email your account has not accessed before.
  • Responses report their cost. Read x-credit-cost and x-credit-remaining from the response headers.

Pagination

On CompanyEnrich, nextCursor is omitted on the last page, and a cursor expires 24 hours after the first page. The Python SDK pages for you with search_all().

Rate limits

Explorium counts each ID in a request toward 200 queries per minute. CompanyEnrich counts requests: a global limit of 500 requests per minute per IP address, and a per-route limit from your plan, reported in x-ratelimit-limit, x-ratelimit-remaining, and x-ratelimit-window. A batch of 50 domains is one request. The SDKs wait and retry on 429. See Rate limits.

Text, URLs, and codes

IDs

Explorium IDs and CompanyEnrich IDs are separate. CompanyEnrich company IDs are UUIDs, and person IDs are 64-bit integers. Store the CompanyEnrich ID next to the Explorium ID, and re-key company data on its domain.

Verify parity on your own data

Measure CompanyEnrich on records you already have before you switch. This makes no Explorium requests, so it works after AgentSource is discontinued.
1

Check your key and plan

Call GET /v2/me. It is free and returns your balance and plan capabilities.
2

Size your searches for free

Send each translated filter to the company or people count endpoint and compare total with Explorium’s total_results.
3

Pick a sample

Take 100 to 200 firmographics records you already stored from Explorium. CompanyEnrich charges 1 credit per company found.
4

Compare the fields you read

Run the script below on the sample. It enriches the domains in batches of 50 and sorts each field into equal, differ, only Explorium, only CompanyEnrich, or both empty. Fields marked lossy in the field mapping differ by definition, so leave them out or compare them by hand.
5

Check searches and people

For each translated search, run the count, then a preview if your plan includes it, then one page. For person lookups, test work emails you already know.
compare.py

Cut over

  • Put the CompanyEnrich path behind a feature flag, and switch it on before Explorium stops serving requests.
  • Store CompanyEnrich IDs next to Explorium IDs, and keep the Explorium values you stored.
  • Log x-credit-cost per call for the first weeks.
  • Remove the Explorium client, keys, and MCP configuration once nothing calls them.

What has no equivalent

Plan a replacement for these, because there is no Explorium to fall back on:

What you can add after migrating

Semantic company search

Search companies with a plain-language description instead of filters.

Similar companies

Find lookalikes of up to 10 seed domains.

Company headcount

Headcount by department, with monthly history.

MCP server

Give AI agents access to the same data.

Common questions

No. Store the CompanyEnrich ID alongside the Explorium ID, and re-key company data on its domain.
No. Send the domain, or the name or a social profile, and the record comes back in the same response.
Firmographics, technographics, funding, workforce trends, and lookalikes have one, with fewer fields. See the endpoint mapping for the rest.
No. CompanyEnrich finds work emails with GET /v2/people/{personId}/email.
Send the same filter to the free count endpoint. Search responses do not include a total.
No. Batch and bulk enrichment work on every plan.
Explorium lowercases place names. CompanyEnrich returns the source casing. Lowercase values yourself if your code compares strings.