Trust & Security
We handle sensitive requirements docs and proprietary plans. Here's exactly how we protect them — and you.
Every workspace requires a signed NDA before any document can be uploaded or processed. The NDA is presented in-product and signed digitally — no PDFs to email back and forth.
A countersigned copy is emailed to you immediately and stored in your workspace. The NDA covers all materials uploaded to Plainframe, all generated plans and artifacts, and all communications during the engagement.
Current NDA version: 1.0
Scope: All workspace data, uploaded documents, generated plans, and communications
All data stored in MongoDB Atlas with AES-256 encryption. Database clusters run in EU-West (Ireland) with automated backups encrypted at the same standard.
All connections use TLS 1.3. HTTP Strict Transport Security (HSTS) is enforced with a one-year max-age, including subdomains.
Sessions use cryptographically random 256-bit tokens, stored server-side with automatic expiry. Cookies are HttpOnly, Secure, SameSite=Lax, and scoped to the application domain.
Passwords are hashed with bcrypt (cost factor 12). Plaintext passwords are never stored or logged. Account lockout triggers after 5 failed attempts.
Uploaded documents are retained for 90 days from the upload date, then permanently deleted via MongoDB TTL indexes. This is automatic — no manual intervention required.
Generated plans, workspace data, and audit logs follow the same 90-day policy. When a workspace is deleted, all associated data is purged.
Every state-changing action is logged: document uploads, plan generation, sharing, settings changes, and authentication events. Logs include timestamp, user ID, workspace ID, action type, IP address, and user agent.
Audit logs are stored immutably and can be provided to workspace owners on request. A self-serve audit viewer is rolling out.
Plainframe uses leading AI APIs — including Anthropic Claude, Google Gemini, and others — for document analysis, requirements extraction, and plan generation. We route across multiple engines for maximum availability: if one provider is unresponsive, requests automatically fail over to another.
All AI providers we use are selected for their API data policies — inputs and outputs are not used for model training. Data is processed to return results and is not retained beyond standard API processing windows.
We do not fine-tune or train any models on your data. Your content is used only to generate outputs for your workspace — not to benefit other customers. The specific engine used for a given request is an implementation detail; your data protections are the same regardless of which provider handles it.
The AI engines are infrastructure we run on — they're not the moat. Plainframe's advantage is the system around them: the structured evidence format, the extraction and planning pipeline, and the validation loops that make outputs consistent and defensible.
Your documents, requirements, and generated plans remain yours. We use your inputs only to generate outputs for your workspace — we don't fine-tune models on your data and we don't reuse your content to benefit other customers.
We improve through operational signals — not by training on your documents. That includes ambiguity rates, plan revisions, scope-change reasons, validation outcomes, and estimate-vs-actual deltas when you run sprints. These signals help us calibrate scoring and planning quality over time without learning from your proprietary content.
Our IP: structured formats, schemas, prompts, scoring rules, and validation workflows
Your IP: uploaded documents, requirements, plans, and artifacts generated for your workspace
Our improvements: derived operational signals (e.g., ambiguity, revisions, validation outcomes) — not document content
Plainframe uses the following third-party services to deliver the product:
Additional AI providers may be added for availability. All providers are vetted for equivalent data protection policies before integration.
Plainframe is a multi-tenant application with strict workspace isolation. Every database query is scoped to the authenticated user's workspace — there is no mechanism for cross-tenant data access.
We do not require access to your internal systems. Planning and analysis happen entirely within Plainframe. Product sprints run in isolated sandbox environments that we provision and hand off to you.
The Plainframe team does not have standing access to customer workspace data. Database access is restricted to infrastructure operations and requires explicit authentication.
Plainframe is built by Variant Work Consulting. Small team, low risk: the work is system-driven, versioned, and exportable — not dependent on tribal knowledge. Real operators review outputs and maintain the delivery workflow, so you get consistency you can trust.
Variant Work Consulting is the contracting entity for all engagements, invoicing, and legal agreements. For procurement, vendor onboarding, or W-8/W-9 documentation, contact us directly.
Contact for vendor details: [email protected]
If your procurement or security team needs additional details, we're happy to answer specific questions or provide documentation.