LetterMate

Twenty letter workflows built around the facts a user needs to provide

Services: Document AI ; AI Workflow Automation
Delivery: Guided legal-letter product, document generation, subscriptions, and administration.

The project reports 12,000+ generated letters and 70% lower cost compared with lawyer-led letter creation.

The product serves users who know what happened but do not know how to turn it into a formal letter. A rent dispute needs different information from a refund request or an employment grievance.

An unrestricted chat interface would leave users responsible for choosing the right prompt and remembering the required facts. LetterMate needed guided forms and letter-specific generation.

The generator covers twenty defined workflows, including rent disputes, deposit returns, parking appeals, refund requests, subscription cancellation, and employment grievances.

Each workflow collects its own fields. For a rent-increase dispute, the generator uses details such as the notice date, current rent, proposed rent, and reason for the dispute. It requests structured body paragraphs rather than returning an unformatted chat response.

The application adds user accounts, saved-letter history, PDF export, subscriptions, checkout, support, and administrative views.

Write prompts around a letter type. The required inputs and output instructions change with the user's task.

Return a document structure. The generator requests JSON containing body paragraphs. Application code parses and formats the response for the document view.

Keep formatting out of the user's hands. Cleanup removes unwanted subject lines and placeholders from the generated body.

Support repeat use. Letter history, billing, and account management are part of the product rather than a separate manual service.

The delivered product gives users a path from a specific issue through a questionnaire to a reviewable letter and export, with 12,000+ letters generated and a reported 70% lower cost compared with lawyer-led letter creation.

The code supports the twenty-workflow scope. It uses prompt-based OpenAI generation; it does not establish proprietary fine-tuning. Generated drafts still require user review and are not a guarantee of a legal outcome.

Next.js 15, React 19, MongoDB, Mongoose; OpenAI with a configurable model defaulting to GPT-4.1 in the generator; React Hook Form, Zod, jsPDF, formatting utilities; Stripe; JWT, bcrypt, and email integrations.

Twenty guided paths to a formal letter.
LetterMate asks letter-specific questions and turns the answers into structured drafts, with document export, history, and subscriptions. The project reports 12,000+ letters generated and 70% lower cost compared with lawyer-led letter creation.

We built LetterMate around twenty distinct letter workflows. The user answers relevant questions, the model returns structured paragraphs, and the product handles the document, history, and billing around it.