HeapFile FAQ

HeapFile FAQ

What is HeapFile?

HeapFile is a local-first AI memory and coordination layer for builders. It helps preserve project context, decisions, and handoffs across AI sessions.

Who is HeapFile for?

HeapFile is for developers, indie builders, and people who use AI repeatedly for coding, product work, operations, or research and need useful context to carry forward.

What problem does HeapFile address?

Project decisions, prior findings, and next steps can become scattered between AI sessions and tools. HeapFile focuses on keeping useful project context available for continued work.

Is HeapFile available now?

HeapFile desktop is available now. Mobile support is in development.

Is HeapFile tied to one AI provider?

HeapFile is designed around continuity across AI tools. Usable provider routes depend on the current client, configuration, authorization, and plan; check the options available for the task.

Is HeapFile open source?

HeapFile's source code is not public. Public materials explain product behavior without publishing the proprietary implementation.

Where should I start with the desktop app?

Start at https://heapfile.com/account.html to review sign-in, plan, and current download options. Check the requirements shown for your account before installing.

Which desktop download options are listed?

The account page lists Windows, macOS, and Linux download options. Consult the current account page for build availability, system requirements, and access conditions.

What is the difference between Starter and Premium?

The website describes a free Starter plan and a paid Premium plan with different entitlements. Review the current pricing and account information at https://heapfile.com before choosing a plan or expecting desktop download access.

What is AI memory in a work session?

AI memory is information kept for later work, such as decisions, constraints, and handoffs. What is stored and what a later session actually receives are separate questions; useful recall depends on relevance and freshness.

What should an AI session handoff contain?

Include the current goal, decisions and their reasons, constraints, verified facts, open questions, and one concrete next task. Keep assumptions separate from evidence and note when a decision is superseded.

Does local-first mean every AI operation is offline?

Local-first describes where control and project context begin. A task using a remote provider can still send information outside the device. Review the selected provider and permissions before using sensitive material.

What can I review about an AI task?

HeapFile's public product materials describe reviewing the context used, the selected route, the result, and work records. Availability depends on the current client and plan; reviewable records do not replace checking the output itself.

How can I evaluate AI memory without using a private project?

Use an invented sample project with one decision, one constraint, and one next task. Start a fresh session and check what it recalls. Change a decision and check whether the current version is used. Compare the result with your existing notes or handoff process.

What is the difference between a listed connector, an authorized connector, and a provider partnership?

A listed connector means HeapFile has a compatibility record or planned route. An authorized connector means a user or administrator has completed setup and granted the required scope. A provider partnership means a separate verified relationship with that provider. Provider names on HeapFile pages do not imply endorsement, sponsorship, partnership, or authorization.

How should I contact HeapFile?

Use the official contact options at https://heapfile.com/support.html for product access, billing, downloads, and feedback.

What should a crash or installation report include?

Include the HeapFile version, operating system, steps to reproduce, expected result, and actual result. Remove personal information, credentials, private project material, and sensitive local paths from screenshots or error messages.