contractiq.app

ContractIQ × SOW Maker

AI that reads requirements and writes contracts and SOWs - so agencies stop bleeding hours on documents.

Timeline
2026
Role
Founder / architect
Team
Solo
Status
building

Technology Stack

Next.jsNext.jsTypeScriptTypeScriptNode.jsNode.jsMongoDBMongoDBLLM pipeline

Overview

ContractIQ x SOW Maker is AI that reads a requirements exchange and writes the documents that usually eat an agency's week - SOWs, BRDs, payment schedules, meeting checklists. One engine, many document types, data-driven rather than hardcoded per format.

It is deliberately architecture-first: the entire system was specified in a master SRS before the first line of code. Every freelance and agency deal drowns in the same documents, written by hand and inconsistently, every single time - ContractIQ turns that repeated manual work into a generated, versioned artifact.

Key Features

  • Requirements → client-ready SOW - An AI-assisted drafting flow turns a requirements conversation into a statement of work a client can sign - already used on real engagements.

  • One engine, many documents - BRDs, payment schedules and meeting checklists come out of the same data-driven template system, not a separate hardcoded generator per format.

  • Spec-first discipline - Database schema, API contracts, security framework and delivery roadmap were written up front - engineering executes against a blueprint, not vibes.

  • Versioned document control - Distribution and access levels on every document - consulting-grade discipline applied to my own product.

Technical Implementation

01Master SRS v2.0

The system is defined by CIQ-SRS-2026-001 - 14 functional modules, each with personas, product pillars and acceptance criteria fixed before implementation. The build is executing that spec, not discovering it.

LayerSpec artifactPurpose
RequirementsSRS v2.0 · 14 modulesPersonas, pillars, acceptance criteria
DataSchema + ER modelDocuments, templates, versions, clients
APIContract-first endpointsStable surface the UI builds against
AILLM drafting pipelineRequirements → SOW / BRD generation
SecurityAccess + distribution levelsVersioned, controlled document release

Design Decisions

Specify before building

For a document engine, ambiguity compounds - a vague SOW schema means every generator downstream is guessing. Writing the SRS first turned the hard thinking into a blueprint, so implementation is mechanical instead of exploratory.

One data-driven engine

Rather than a bespoke code path per document type, ContractIQ treats every artifact as data over a shared engine. Adding a new document format becomes configuration, not a new subsystem.

Want the full walkthrough - architecture, decisions, war stories?

ask me about it