Beta. AgentSource v2 is in beta. Details may still change before GA.
Structural changes
New capabilities
- Research (GenAI) endpoints — custom AI-powered research over businesses or prospects, with a caller-defined
output_schemaand optional real-time web grounding. - Async job infrastructure — shared job submission, status, and result retrieval behind every
.../jobendpoint, with optional email notifications on job success or failure. - Multi-webhook, tenant-scoped webhook management — up to 10 webhooks per tenant, addressed by a server-generated
webhook_id, each with its own rotatable secret; enrollments bind to a webhook explicitly. Replaces v1’s single partner-level webhook and its silent overrides. See Webhooks. - Unified contact-availability filter — prospects fetch and stats gain
has_contact_details(email,phone,email_or_phone,email_and_phone), including the previously impossible “at least one of email or phone”. - Contact information waterfall — the async contact endpoint can fall back to a verified external provider for prospects Explorium data cannot enrich, with per-result source tagging.
Signal fixes
Misspelled field names corrected in v2:Renamed
Removed fields
entity_id— no longer returned for enrichmentsdisplay_name— removed from the businesslinkedin_postsenrichment response- Address signals removed from the
company_ratings_by_employeesenrichment:company_citycompany_country_codecompany_regioncompany_streetcompany_zip_code
Removed filters
has_email and has_phone_number are removed from the prospects fetch and stats filters. Use has_contact_details instead:
email_or_phone — at least one of email or phone — is new in v2 and had no v1 equivalent.
Changed defaults and parameters
- The
sizeparameter is removed from pagination in fetch. The default stays 60K. include_operating_locationsin business fetch now defaults tofalse.
New limits
Coming in Phase 2
max_per_company support. See the release plan.
Next
Migrating from v1
Endpoint mapping and the v1 deprecation timeline