The core local continuity system.

Continuity Engine Basic is the core local continuity system for serious AI work. It preserves local chat records, supports transcript search, lets users mark important highlights, and builds controlled context packages so work can continue in a fresh chat when the old one gets too long.

Long AI projects drift, slow down, and eventually need to move. Without a continuity layer, users spend their lives re-explaining old decisions, copying fragments, and hunting for the one sentence that mattered. Continuity Engine gives the project a local memory spine.

The local archive can include recreated chat transcripts, independent HTML Archive backups, project/session records, selected highlights, images, files where available, summaries, update packages, and project backup material. Each imported chat can be preserved as a local HTML record with linked assets when capturable, so the user keeps a usable copy outside the aging browser chat.

The standard Continuity Engine workflow is designed around local storage. Betweenware does not need your private transcript archive in order for the local workflow to operate.

No. Basic is not built around requiring a customer AI API key. API-backed search and merge behavior belongs to optional PRO functionality.

ChatGPT is the primary full workflow target. Claude and Gemini browser workflows are supported through the same local continuity layer, with documented manual or platform-specific handoff steps where their interfaces require them.

Start the local Continuity Engine first, then open a supported chat page and click the Chat Health floater in the lower-right corner of the chat. When the extension detects that the Engine is running, the floater switches to the Continuity Engine icon and opens the CE sidebar. If the Engine is installed but not running, the sidebar will not open because the extension stays in Health Meter or Schrodinger's Chat mode.

It is the recreated local record of the chat session. The point is to keep the original work available so you can search, review, quote, preserve decisions, and build better handoffs instead of relying only on memory or vague summaries.

HTML Archive creates an independent local HTML copy each time a chat is imported into Continuity Engine. It is saved with linked image and file assets when available, so the user has a permanent transcript that can be opened and searched outside the Engine.

This matters for data security and long-term access: even if the user stops using Continuity Engine later, the captured chat backup remains on their hard drive as an HTML transcript rather than an Engine-only database record, PDF, or summary. Safe and searchable. Local and identical.

Transcript search lets you find exact phrases, decisions, requirements, links, or turning points across a project archive. For long-running work, this is often the difference between continuing the actual project and accidentally rebuilding a slightly warped copy of it.

Highlights are user-selected pieces of context that should not get buried. They can include important text, decisions, formulas, images, file references, and other high-value material.

One-shot highlights are meant to travel into the next update and then retire. Pinned highlights are meant to keep traveling across ongoing chats until the user unpins them. One-shot is a flare. Pinned is a lighthouse.

Continuity Engine can preserve image and file targets when they are available through the browser interface or active download links. Download links can expire, so users should not assume that old chat downloads remain available forever.

Yes. The project/session structure is meant to separate work by project so multiple long-running efforts do not melt into one soup.

Yes. You can import a chat, keep working in that same chat, then import it again later. Continuity Engine detects what was already saved and stores only the new material as a new session in the same project.

This is useful if you want to save your work before using AI Browser Reset, continue if the chat improves, then import again before building a final update package.

A new-chat update package is the context bundle used to start or continue work in a fresh AI chat. It is designed to be compact enough to fit into a new prompt but rich enough to preserve the project's real decisions, goals, and selected evidence.

The Basic update package combines a deterministic capsule, an AI-generated summary layer, and a Library document containing summary highlights and user-selected highlights. The capsule preserves structure and original language, the summary captures meaning, and the Library carries selected evidence.

Yes. Claude and Gemini do not allow Continuity Engine to automate every prompt-box step the same way ChatGPT does, so the result is the same, but a few sends and downloads must be done manually.

For Claude and Gemini, import the HTML transcript, choose any highlights, click Summary, then manually press the AI platform's Send button. When the AI generates the Summary document, download it to the Windows Downloads folder. Open a new Claude or Gemini chat, click Update New in Continuity Engine, then before sending drag the Summary file into the new prompt next to the Deterministic Capsule and Library. Click Send to complete the handoff.

It means a controlled compression path that preserves structured transcript context more predictably than asking an AI to freestyle a summary. It is not magic perfect memory, but it is much less squishy than a loose summary alone.

The update token target can be selected by the user, including common targets like 5K, 10K, and 20K, plus a user-designated larger range where supported. The right size depends on model limits and how much context the next chat needs.

Yes. Basic is designed to carry forward a project across multiple sessions. Multi-session updates help when the project has grown across several old chats rather than one long conversation.

Yes, and that is a feature, not a defect. CE helps preserve and structure the work, but the user still chooses what matters, when to update, and how much context to carry forward.

Check that the Engine is running, that the browser extension is enabled, and that the local health endpoint is reachable where the manual says it should be. If it still fails, email support with your OS, browser, Engine version, extension version, and a screenshot of the status.

Check the Updates page first. If the issue is not listed, email support. Browser-interface changes can affect capture or injection, but already saved local records should remain available.

Some images, generated files, and download links are exposed differently by browser AI platforms. If a link has expired or is hidden behind a changing interface, CE may not be able to retrieve it. Save critical files directly when they matter.

The product supports project backup and archiving workflows. Use the manual's drive-location instructions carefully, and keep backup copies before moving important archives.

For product support, email support@betweenware.ai. For general pre-purchase questions, email contact@betweenware.ai.