prose → oscal
AI-assisted extraction
The assistant extracts structure from prose with provenance and confidence. It never invents OSCAL content — missing fields are flagged as gaps.
Loaded deterministic baseline. Click Extract to call Lovable AI.
Target model
Prose input
paste a control narrative, POA&M, or system description
Extracted OSCAL draft
SSP · oscal 1.1.2
ai user needs info
Assembled implemented-requirement (preview)
json · will be embedded into control-implementation.implemented-requirements
1{2 "uuid": "<assign on save>",3 "control-id": "ac-2",4 "statements": [5 {6 "statement-id": "ac-2_smt.a",7 "uuid": "<assign on save>",8 "by-components": [9 {10 "component-uuid": "<select component>",11 "uuid": "<assign on save>",12 "description": "Acme operates a SaaS platform on AWS GovCloud. The platform supports three account\ntypes: privileged administrators, service accounts used by automation, and end-user\naccounts for agency staff. Account types are defined in our IAM policy and enforced\nat the API gateway component.",13 "props": [14 {15 "name": "account-types",16 "value": "privileged administrators"17 },18 {19 "name": "approval-process",20 "value": "Manager approval via ticketing system prior to provisioning."21 }22 ]23 }24 ]25 }26 ]27}Schema check before save
ssp · 1.1.2
2 schema errors1 warnings
- $.system-security-plan.system-characteristics.system-idsSSPs must include at least one system identifier (e.g. a FedRAMP package ID) under system-characteristics.system-ids.
- $.system-security-plan.system-implementation.componentsAn SSP must list at least one component under system-implementation.components.
Draft cannot be saved as valid until all schema errors are resolved. Fill in the system-characteristics fields and at least one component to clear them.