SolScribe 2.0 changes what happens after a recording is transcribed. Brain checks each new fact against what it already holds. Summaries and tasks know who said what, long transcripts are processed in parts and Home shows what you are waiting on from other people.
2.0 went live on SolScribe Cloud on November 2, and the follow-up releases through 2.0.0.3 went out on November 7.
Brain updates existing entries
Before 2.0, a fact restated in a new meeting became a new entry. Now Brain compares each new fact with similar entries you already have. A restatement refines the existing entry and keeps the old wording in its history. A fact that explicitly replaces another marks the old one superseded, with a link to its replacement. A possible contradiction waits for your review.
Superseded entries are kept, but they drop out of search, links and digests. Brain also skips overheard conversations and attendance, and records each decision with its basis: decided, proposed or observed.
Who said what
Brain, summaries and reports now see speaker turns, your vocabulary, the people and projects you track and the day the recording was made. Your own voice profile shows as Owner, and “next Friday” resolves against the recording date.
That context lets SolScribe tell a request you made of a named person from a promise someone made you. Both show in Tasks, as “Asked of Mira” or “Mira (tentative)”. Your own hedged ideas do not become tasks.
Open loops
Home and Brain now show open loops: promises someone else made you and tasks you marked as waiting. A loop is open, stale after 14 quiet days, expired after 60 or resolved. Home shows at most three.
Long recordings
Recordings over 12,000 words now run Brain and the structured summary in parts and merge the results, so the second half of a long meeting is no longer skipped. Processing stays within coverage and time limits, up to six parts per recording. A very long transcript can exceed them. When coverage is incomplete, SolScribe flags it, and a summary that runs out of time is saved as partial and says which parts it covers.
Summary length and templates
Structured summaries scale with the recording, so a short clip gets a short overview and a long meeting gets more key points and decisions. They can also follow a template: a name and up to eight sections, created in the web app. Your default template, or one you pick, adds its sections after the overview. A section with nothing to report stays empty rather than filled in.
You can add your own notes to a recording. Notes shape the summary and help focus Brain, but Brain facts still need transcript evidence, and a task supported only by your notes waits for your review.
Playback timestamps and the Daily digest
New tasks, task proposals and Brain entries play from where they were said when the supporting quote can be located in the transcript. Existing items need reprocessing to get a timestamp. Transcripts now mark low-confidence words and let you step through them.
The Daily digest has cited sections for highlights, decisions and open questions, each naming the recordings it rests on. It now shows on Home as well as Brain, one day at a time.
Editing recordings
Recordings can be split, clipped and merged. Each previews the result and the originals are kept.
Review defaults and limits
- New accounts review extracted tasks before they become active. Existing task workspaces keep their mode and can switch review on from Tasks.
- Tasks from recordings in a language other than English always wait for review, because the checks that separate a promise from a request only understand English.
- Summaries, tasks and Brain entries are AI-generated. Check them against the recording before relying on them.
More at solscribe.ai.