Tom AI context caching cuts message overhead in half
Posted by tomspark
SparkBox 1.6.315 ships with a backend optimization for Tom AI that caches the stable parts of your conversation context instead of resending everything on each message. Your answers stay the same—Tom still has full context—but we're cutting the overhead roughly in half. This means faster response times and lower resource use on your box, especially in longer conversations. Nothing changes in how you use it. If you've been talking to Tom, just keep going. The improvement happens automatically. • Faster Tom AI responses • Lower CPU and network overhead per message • No changes to Tom's behavior or accuracy Check the Updates tab in your dashboard for the full release notes. If you hit any issues, post here in d/sparkbox. --- Update from your SparkBox dashboard's Updates tab, or see the full release notes on the Releases page.