AgentSource v2 is now available in beta. It is the biggest revision to Explorium’s API since launch, built for one problem specifically: AI agents that need to run larger, more reliable data workflows without babysitting them.

    If you are already on AgentSource v1, nothing changes today. v1 keeps running in parallel for six months, so there is no forced cutover and no broken integration on launch day.

    What’s new in AgentSource v2

    v2 is a genuine redesign, not a version bump. Four changes matter most:

    • Async jobs at scale. Run workflows against up to 10,000 records per job, with job runtimes up to 24 hours and optional email notifications when a job completes. Every enrichment now offers both a synchronous .../enrich call and an asynchronous .../job call, so you choose the mode that fits the workload instead of building your own polling layer around a sync-only endpoint. The async guide walks through submitting a job, polling its status and collecting results.
    • A contact waterfall for better coverage. Async contact enrichment can now fall back to verified external providers when the primary source comes up empty, with the source tagged on every result. A new has_contact_details filter (email, phone, email_and_phone, or email_or_phone) replaces the old has_email and has_phone_number flags with one consistent parameter. Waterfall results are charged at a higher rate than Explorium’s own contact data — externally sourced phone numbers cost 15 credits against 5 for an Explorium phone, while emails stay at 2 either way — so check the credit charging details before turning it on for a large job.
    • AI-powered research endpoints. A new research capability runs custom AI research over businesses or prospects against a caller-defined output schema, with optional real-time web grounding, so an agent can ask a genuinely open-ended question instead of only pulling structured fields.
    • A cleaner, unified API surface. Every endpoint now lives under /v2. Separate bulk routes are gone: one endpoint now accepts either a single ID or a list. One /v2/autocomplete replaces the old per-entity autocomplete routes. Seven previously misspelled fields are fixed (for example, professions_email is now professional_email, and founding_stage is now funding_stage).

    Alongside those, webhooks got a real upgrade: tenants can now register up to 10 webhooks, each with its own secret, replacing v1’s single partner-level webhook. And a max_per_company control for prospect results is planned for a follow-up release.

    Why this matters if you’re building on top of Explorium

    Most of what changed maps directly onto the failure points teams hit when they try to run enrichment or research at agent scale rather than at dashboard scale: a sync-only API that times out on large batches, a single webhook that can’t route to more than one downstream system, inconsistent contact filters across endpoints, and a bulk/non-bulk split that forces two code paths for the same job. v2 removes those constraints without asking you to redesign your workflow, since v1 stays live in parallel while you migrate at your own pace.

    A new $29.99 package

    Alongside AgentSource v2, we also launched a new $29.99 package, giving teams a lower-commitment way to get started with Explorium before scaling up to a larger plan.

    Migrating from v1

    There is no hard cutover. v1 will continue running for six months, which gives existing integrations a full runway to move over. The migration guide maps every v1 endpoint to its v2 equivalent and sets out the full deprecation timeline.

    Get started

    AgentSource v2 is in beta today. Read the full overview and endpoint reference at the Explorium API v2 documentation, or see everything that changed in the what’s new in v2 page.

    FAQs