Doorframe

Doorframe is an open-source, local-first requirements traceability and review tool. It turns requirements exports, Jira work items, and test results into a traceability gap report you can use before a review.

Doorframe is useful without AI. The main product artifact is the HTML traceability gap report.

Doorframe runs locally by default and does not send imported project data to any external service.

Install And Run

Most users should use the doorframe npm package or Docker. Clone the repo only when contributing, auditing, or changing Doorframe.

Try Doorframe without installing:

npx doorframe demo

Run the local web app with npm:

npx doorframe serve

Open the printed local URL, normally http://localhost:3000.

Generate a report from local exports:

npx doorframe analyze \
  --requirements requirements.csv \
  --jira jira.csv \
  --junit test-results.xml \
  --out doorframe-report.html

Install globally:

npm install -g doorframe
doorframe --help
doorframe serve

Run the local web app with Docker:

docker run -p 3000:3000 -v doorframe-data:/data ghcr.io/vtboyarc/doorframe:0.1.3

Run MCP:

npx doorframe mcp --project ./doorframe.sqlite

See docs/install.md, docs/docker.md, and docs/company-installation.md.

What Doorframe Is

• A local review tool for requirements exports.
• A local web app, npm package, Docker setup, command-line runner, and report generator.
• A deterministic analyzer for traceability gaps, weak wording, duplicate candidates, failed tests, stale links, and closed work without passing verification.
• A way to compare two requirements baselines before a review.
• An optional read-only MCP server for local project databases.

What Doorframe Is Not

• Not a DOORS, Jama, Polarion, Jira, or test-management replacement.
• Not a SaaS.
• Not an AI requirement generator.
• Not approved for classified data.
• Not a compliance certification.
• Not a substitute for official program, product, quality, security, or audit processes.

Ten-Minute Demo

npx doorframe demo

Open doorframe-report.html in a browser. The report is offline HTML and can be printed to PDF. The demo also writes doorframe-baseline-diff.html.

The demo project is fictional. It intentionally includes weak wording, missing trace links, missing passing verification, closed work without passing tests, duplicate requirement candidates, and a baseline timing change.

Baseline Diff Demo

npx doorframe diff \
  --baseline-a ./examples/falcon-telemetry-gateway/sample-requirements-baseline-a.csv \
  --baseline-b ./examples/falcon-telemetry-gateway/sample-requirements-baseline-b.csv \
  --out ./doorframe-baseline-diff.html

This detects REQ-014 changing from a 5 second processing threshold to a 2 second threshold, plus one added and one deleted requirement.

Web App

With Docker:

docker run -p 3000:3000 -v doorframe-data:/data ghcr.io/vtboyarc/doorframe:0.1.3

From source:

npm run dev

Open http://localhost:3000, create a project, and click Load Demo Project. Doorframe loads the fictional Falcon Telemetry Gateway sample and generates findings, trace links, and reports from local data.

Docker

docker compose up

Open http://localhost:3000. Docker stores project data in the doorframe-data volume mounted at /data.

MCP Server

Doorframe includes an optional read-only stdio MCP server for local Doorframe SQLite project databases.

npx doorframe mcp --project ./.doorframe/doorframe.sqlite

The MCP server is not an AI client and does not call AI providers. Data returned by MCP may enter the connected AI client's context.

Inputs

• Requirements CSV
• ReqIF and ReqIFZ
• Jira CSV
• JUnit XML from common CI/test runners

CSV and Jira imports preserve raw attributes where practical. Requirement IDs are extracted from configured patterns in summaries, descriptions, labels, custom fields, test names, class names, suite names, and failure messages.

Sample Report

Generate the demo report with the npm command above. A screenshot can be added after the first release artifact is published.

Project Structure

apps/web            Next.js UI, API routes, and SQLite persistence
apps/cli            Doorframe CLI
apps/mcp-server     Read-only stdio MCP server for local project databases
packages/core       Shared types, schemas, rulesets, baselines, and utilities
packages/parsers    CSV, JUnit XML, ReqIF, ReqIFZ, and ID extraction
packages/analyzers  Deterministic traceability rules and baseline diff
packages/reporting  HTML, Markdown, JSON, CSV, and baseline diff reports
packages/storage    Shared project-data assembly and read-only SQLite loading
examples            Fictional sample data
docs                Project docs
docker              Docker image definition

Security And Privacy

Doorframe has no telemetry in the MVP. It does not phone home, require accounts, or call AI providers by default. Reports are generated without external scripts, fonts, images, or CDN assets.

See SECURITY.md, docs/security-model.md, docs/threat-model.md, docs/no-telemetry.md, and docs/data-handling.md.

Roadmap

Near-term work for v0.1.x focuses on better import resilience, clearer reports, baseline workflows, packaged demo artifacts, and user feedback from real review workflows. Optional AI review can come later, but Doorframe must remain useful without AI.

Contributing

Clone the repo only when contributing, auditing, or changing Doorframe.

git clone https://github.com/vtboyarc/Doorframe.git
cd Doorframe
npm install
npm run build
npm test

See CONTRIBUTING.md. The repository currently uses the existing MIT license in LICENSE.