Start here

Petria is a petition and application engine, not a general immigration research chat. The two objects you work with are matters and jobs. A matter is one client case, tied to a petition or application type and a jurisdiction. A job is one run through the pipeline against that matter: a full draft job, an RFE or RFR response job, or a section regeneration on Enterprise. You open a matter, attach exhibits, submit a job, and download Word deliverables when it finishes. There is no separate export step and no chat window to manage.

Behind that simple flow is a fixed pipeline every full draft job runs. PII interception tokenizes names, dates, and other identifiers before anything else happens. Evidence normalization translates or preserves each exhibit against the matter's working language. Extraction pulls facts and typed claims out of the tokenized record. Drafting writes the petition section by section against a researched template plus your firm's voice. Citation extraction and verification check every claim the draft made against your actual exhibits, and redraft fixes what verification flagged. On Pro and Enterprise a pre-filing gap analysis runs after that first pass, and Enterprise can run a second, targeted redraft loop against the gaps it finds. De-tokenization restores real names and dates only in the final delivered documents, never in an intermediate artifact a drafting model reads.

Most firms start onboarding with a handful of voice samples (this step is optional; see Set up your firm voice), then create one real matter, upload its exhibits, and run a single full job so the team can see the draft, the verification report, and, on Pro and Enterprise, the gap analysis together before deciding how the product fits their workflow. Reviewing all three side by side, rather than skimming just the draft, is what tells you fastest whether Petria is grounding claims in your record the way you expect.

When a job finishes, everything lands in the matter panel automatically: the draft, the verification report, the exhibit list, and any response-job filing packet. Full draft jobs are metered per firm per calendar month (Standard: 5, Pro: 10, Enterprise: custom), and RFE and RFR response jobs draw from a separate monthly bucket that is Pro and Enterprise only.

Tier controls three things: which layers Petria learns from your samples (voice only on Standard, plus emphasis on Pro, plus an approved structure on Enterprise), whether RFE and RFR response jobs are available at all (Pro and Enterprise only), and whether a second, targeted redraft loop runs against the gap analysis before delivery (Enterprise only). All three tiers run the same core pipeline and the same verification step. The difference is what gets layered on top, and how much of the process your firm has locked in versus what Petria decides fresh on every job.