Mulch
Back to the app →

Changelog

What's changed

A running log of updates to Mulch during the private beta, newest first. Since this is a proof of concept moving quickly, expect this list to grow often — check back if you're curious what's new.

  • Learning edges: one deliberate edge you carry into every session, plus any number of held edges that are alive but not your current focus, and settled ones kept as history. Manage them in Learning Edges, from Clinician Profile or the Growth Reflections sidebar. Your current learning objective became your deliberate edge.
  • Deliberate Practice reflections now ask how your deliberate edge showed up and whether the session touched a held edge, and can stay with a held edge instead of steering back. After a reflection saves, you can attach it to the edge it was about.
  • When Mulch suggests a new learning objective, it now asks before saving it.
  • Session self-reports: when you start a new note, Mulch turns the previous note conversation into a short record of what you said about yourself (your state, reactions, what you tried and why), with anything that could identify the client removed. Read, edit, or delete them in Reflection History; Growth Reflections reads your latest 60.
  • Saved reflections can now be deleted from Reflection History.
  • Corrected the help page, beta page, and privacy policy on how long note conversations are kept.
  • Added Terms of Service and Privacy Policy pages, linked from the signup and login screens and from Settings.
  • Deliberate Practice and Growth Reflections now save automatically as soon as you navigate elsewhere in the app, close the tab, or log out — not only when the closing message says so — fixing cases where a reflection could go unsaved without any indication.
  • Added a "Save Reflection" button right in the chat box for both reflection types, so you don't have to look for it in the sidebar.
  • Behind-the-scenes cleanup to how note-format instructions are organized — no visible changes.
  • Fixed a login bug where an account created with different capitalization or extra spaces in the email wouldn't log in, even with the correct password.
  • Note intake now asks about extended sessions and anything else to add in the same message as confirming intervention labels, instead of as a separate step later.
  • Deliberate Practice reflections now open by asking how your focus area showed up, rather than guessing at a specific moment from the note.
  • Deliberate Practice reflections are shorter and more focused — one question per turn.
  • Ending a Deliberate Practice reflection (e.g. "that's enough," "I'm good") now saves it in that same reply.
  • Added Weekly Learning — once a week, jot down one case that stood out and one thing worth remembering. No AI involved.
  • Fixed a bug where opening Clinician Profile or Note Settings and saving too quickly, before your existing values had loaded, could overwrite them with blanks.
  • Added an "include diagnosis" checkbox to Note Settings — off by default, so notes don't name a diagnosis unless you opt in.
  • Added basic, privacy-respecting analytics so we can see which parts of the app are actually getting used during the beta.
  • Feedback sent through the Support button is now saved on our end in addition to being emailed, so nothing gets lost.
  • Added a Support button and this Help & FAQ page.
  • Added "Reflect on This Session" — a way to start straight into a Deliberate Practice reflection without writing a note first.
  • Reorganized the sidebar into clearer sections and cleaned up the color scheme.
  • Notes and reflections now use your saved timezone instead of the server's.
  • Added Reflection History, so you can browse, reread, and edit past Deliberate Practice reflections.
  • Rebranded to "Mulch."
  • Smoothed out account creation and beta signup emails.
  • Fixed a bug where dictating quickly could duplicate a message.
  • Added invite-based self-service signup.
  • Note and Growth Reflections conversations are now ephemeral by design — nothing lingers beyond what you choose to save.
  • Added a clinical decision-making guardrail and a disclosure modal about what the AI is and isn't for.