HeapFileAI

Pricing path

Start free. Upgrade when AI context becomes shared, repeated, or expensive.

Free proves local memory, project recall, MCP setup, and local AI context. Premium adds cloud continuity, mobile handoff, smart connectors, team workspaces, savings reports, and managed automation. Bring your own inference; HeapFile sells durable context, handoff, and governance around it.

Upgrade logic

Make the upgrade feel obvious, not forced.

HeapFile should show Premium when the user's work crosses a boundary: repeated context rebuilds, connector evidence, cross-device handoff, team coordination, or background jobs.

Free Local proof
  • Local memory and recall.
  • Project memory and MCP setup.
  • Local git, local files, and Ollama-aware setup.
  • Manual client setup.
  • Personal workflows and feedback.
  • No managed cloud sync, team workspace, or background workers.
Premium Savings path
  • Cloud sync and cross-device handoff.
  • Smart connectors for authorized evidence.
  • Savings and usage reports for local versus hosted AI work.
  • Team workspaces and managed automation.

Billing policy

Manage cancellation and request refund review.

Premium is still delivered as gated services come online. If checkout, entitlement, cancellation, or paid feature access does not work as described, use the refund and support paths.

  1. Refunds. Review the refund policy.
  2. Data controls. Review retention, export, and deletion.
  3. Terms. Review the terms before purchasing.
Individual

Power users

Upgrade when one person needs continuity across devices, AI clients, or source systems.

Team

Shared work

Upgrade when more than one person needs shared context, repeatable setup, roles, and admin visibility.

Enterprise

Governed rollout

Use an enterprise path for security review, retention requirements, audit exports, and custom onboarding.

Data center

Bring your own inference

Use HeapFile beside Ollama, vLLM, NVIDIA NIM, or another local model service to preserve evidence, policy, and spend visibility around the compute you already run.