Alloy

Connect an Alloy account by saving its token/secret pair, after checking it actually works against the chosen environment (sandbox or production).

Disconnect an Alloy account: deletes the saved token/secret. Nothing in Alloy itself is changed.

List the connected Alloy accounts.

Create a Person entity in Alloy from applicant field data (name, DOB, SSN/tax id, address, email, phone, etc.), the object every KYC evaluation attaches to.

Create a Business entity in Alloy from company field data (legal name, EIN/tax id, formation state/country, address, etc.), the object every KYB evaluation attaches to.

Read one Alloy entity (Person or Business) in full, including its evaluation/journey history summary.

Merge a duplicate Alloy entity into a surviving entity -- combines their evaluation/journey history under the surviving entity token.

Add a free-text internal note to an Alloy entity's record -- e.g. context from a manual review or a customer support interaction.

Submit outcome feedback on an entity (e.g. confirmed fraud, false positive) back into Alloy -- feeds Alloy's own model tuning and your account's audit trail.

List the groups (custom entity groupings, e.g. a household or a linked-accounts ring) an Alloy entity belongs to.

Create a Journey Application in Alloy -- runs an applicant through a configured Journey (multi-step KYC/KYB/fraud/credit decisioning workflow) and returns the outcome (Approved/Denied/Manual Review).

Read one Journey Application in full -- its current status, decisioning outcome, and every node's evaluation result.

List Journey Applications for one configured Journey, optionally filtered by outcome status.

Rerun an existing Journey Application through its Journey again -- e.g. after new documents were supplied, or a data source was temporarily unavailable.

Submit a manual review decision on a Journey Application currently in Manual Review status -- approve or deny it as a human reviewer.

Read the manual review token generated for a specific Journey Application alert -- used to link a reviewer directly into Alloy's manual review workspace.

Read a Journey's configured field schema -- every field name/type an application must supply, and every branch/version defined for it. Call before create_journey_application to know exactly what application_data should contain.

Update a single decisioning node's result on an active Journey Application -- e.g. correct an automated document-verification node's outcome after a human review.

Create a Journey Batch -- submit many applications through the same Journey in one call (e.g. a portfolio re-screen or a bulk onboarding import).

Read one Journey Batch's processing status and per-application results.

Run a direct (non-Journey) Evaluation against an Alloy Workflow -- Alloy's original single-step decisioning primitive, still available alongside Journeys for accounts configured that way.

Read one Evaluation's outcome in full -- its decisioning result, triggered rules, and any manual review status.

Run a Portfolio Evaluation -- Alloy's book-of-business-level risk assessment across a set of existing entities, e.g. periodic re-screening of an entire loan portfolio against updated sanctions/watchlist data.

Read one Portfolio Evaluation's aggregated results in full.

Upload an identity or supporting document (e.g. driver's license, business formation certificate) to Alloy for an entity, as a base64-encoded file. Alloy can then run document-verification checks on it via a Journey.

Read one uploaded document's metadata and verification status in full.

Update an existing document's metadata (e.g. correct its document_type classification).

List all documents uploaded for one Alloy entity.

Send a monitoring event (login, transaction, address change, phone change, etc.) for an existing entity into Alloy's Events API -- the core mechanism for continuous post-onboarding risk monitoring, separate from the initial Journey/Evaluation.

List monitoring events recorded for one Alloy entity, optionally filtered by event type -- the ongoing-monitoring history that can itself trigger new alerts/cases.

Open a new Case in Alloy -- a case management record used to track hands-on investigation of an entity/application beyond automated decisioning (e.g. escalated fraud suspicion).

Read one Case in full -- its status, assignee, linked entity, and recorded evidence/work log.

List Cases in the connected Alloy account, optionally filtered by status and/or assignee.

Update a Case's status, priority, and/or assignee.

Attach a piece of evidence (a note, a reference URL, or supporting text) to an existing Case's investigation trail.

Log a unit of investigator work (a summary of what was done, and time spent) against a Case.

Open a new Investigation in Alloy -- a deeper, typed compliance investigation (e.g. AML/SAR-adjacent) distinct from a routine Case, with its own alerts and lifecycle.

Read one Investigation in full -- its type, status, linked entity, and recorded alerts.

List Investigations in the connected Alloy account, optionally filtered by status and/or investigation type.

Update an Investigation's status, title, and/or description.

Archive a closed Investigation, removing it from the active queue without deleting its history.

List the Investigation types configured on the connected Alloy account (e.g. AML review, fraud escalation) -- the values create_investigation's investigation_type accepts.

Create a new versioned Custom List in Alloy (e.g. an allowlist/denylist of emails, SSNs, or device fingerprints) that a Journey's rules can check entities against.

Read one Custom List's metadata -- its name, type, and current version.

List Custom Lists configured on the connected Alloy account.

Add one value (e.g. an email, SSN, or device id) to a Custom List, creating a new version of it.

Remove one value from a Custom List, creating a new version of it.

List the values currently in a Custom List, at its current or a specific past version.

Create a new named version snapshot of a Custom List's current contents (e.g. before a bulk update), so past states remain auditable.

List the Published Attributes configured on the connected Alloy account -- the derived/enrichment fields (e.g. risk scores, verification flags) Alloy computes and exposes per entity.

Read the current Published Attribute values computed for one entity -- e.g. its latest risk score or verification flags.

List items in Alloy's manual review queue, optionally filtered by status.

Read one manual review item in full -- the entity/application it belongs to, and what triggered it.

Submit a human reviewer's decision on a pending review item.

Register a bank account for an entity in Alloy, so it can be verified and its transactions monitored.

Read one registered bank account in full.

List bank accounts registered for one entity.

Record a transaction against a registered bank account, for AML/fraud transaction monitoring.

List transactions recorded against one bank account.

Create a new Group in Alloy for clustering related entities together (e.g. a household, or a business and its beneficial owners).

Add an entity to an existing Group, with an optional role label (e.g. 'primary', 'co-applicant', 'beneficial_owner').

Remove an entity from a Group.

Read one Group in full, including its member entities and their roles.

List Groups configured on the connected Alloy account.

Subscribe to an Alloy event (e.g. journey application decisioned, case status changed) -- Alloy will POST to your endpoint URL as it happens.

List webhook subscriptions configured on the connected Alloy account.

Permanently remove a webhook subscription. Cannot be undone.

List the configured Parameters (account-level constants Journeys/Workflows can reference, e.g. score thresholds) on the connected Alloy account.

Build one aggregated health report across the connected Alloy account: open Cases, open Investigations, pending manual reviews, and samples of recently denied/manual-review applications -- a single-call compliance-ops snapshot instead of stitching together five separate list calls.

Value-add report: read the full pending manual review queue as one flat list of rows (review token, entity/application token, review type, status, created date) instead of paging through list_reviews and cross-referencing each token manually.