Back to Updates

Scheduled Summaries and a rebuilt local LLM engine

Product

The friction-removal release — Scheduled Summaries, Modality Focus, a rebuilt local LLM engine, and expanded BYOK.

5.1.0 removes friction. It puts summaries on autopilot with Scheduled Summaries, adds Modality Focus for capture-aware querying, rebuilds the local LLM engine from the ground up — no more Ollama — and expands BYOK to OpenAI, Google AI Studio, and Anthropic.

What you can do now that you couldn't before

Put summaries on autopilot

Schedule any summary to run automatically — daily, weekly, or on a custom cadence. Morning briefs generate before you open Pieces, and weekly recaps arrive every Friday afternoon.

Query by how you captured it

Modality Focus lets you filter chat prompts by clipboard, audio, or vision — narrowing results to the right capture type when you remember how you encountered something but not the exact content.

Run local models without Ollama

A completely rebuilt local LLM engine removes the Ollama dependency, with fewer setup issues, better stability, and access to newer models.

Bring your own API keys

Full BYOK support for OpenAI, Google AI Studio, and Anthropic joins existing AWS Bedrock and Azure support.

Who benefits most

  • Anyone with recurring reporting needs — standups, sprint reviews, client updates: Scheduled Summaries.
  • Anyone who wants precision in how they search their memory: Modality Focus.
  • Privacy-focused users and teams running models on-device: the local LLM overhaul.
  • Enterprise teams with existing API agreements: expanded BYOK.