Skip to content

v0.2.1

Choose a tag to compare

@github-actions github-actions released this 27 Oct 09:02
· 12 commits to main since this release
1072c30

Features

  • web: fit the interface to a phone (#41) (#225)
  • ask: ask up to three questions in one call, with options (#64) (#224)
  • usage: a Usage screen with a 30-day chart, totals, cache and speed (#111) (#222)
  • web: polish Memory and Mind — one-box entries, a steady switch, tier as an icon (#110) (#221)
  • servers: show every server as a box, switch tools off, add and edit servers (#108) (#220)
  • permissions: make, change and remove rules, and keep them (#109) (#218)
  • web: hotkeys — one keymap, a command palette, find, and a Hotkeys screen (#112) (#199)
  • web: edit and save a published file, and a menu for the chat itself (#112, part 1) (#198)
  • web: every component is the theme's own (#194)
  • web: UI rework on the Heaven theme (#182) (#193)
  • chats: name a chat from a short model summary of its first message (#135)
  • providers: report whether an endpoint's template can render a tool declaration (#145) (#146)
  • web,core: switch a skill off, skill editor (#159)
  • web: collapsible providers and model rows, max output tokens, accepts-images (#156)
  • core,web: account screen with name, avatar and email (#149)
  • web: settings shell with grouped nav, General screen, account card (#148)
  • web,mcp: draw mermaid artifacts (#73) (#125)
  • web: UI primitives — Select everywhere, one modal shell, a custom checkbox (#127)
  • web: a loading pattern, and a screen for when she is not answering (#116)
  • web: show per-message generated token count (#95)
  • web,core: a model configurator (#69) (#94)
  • web: stand her silhouette behind the start and chat screens (#76)
  • a registered model may carry request options (ADR 18) (#68)
  • export a chat as Markdown (#52) (#62)
  • named models per provider, with kind and logo (#49)
  • web: animate the waiting ocellus between feather and eye (#48)
  • web: open a long reasoning trace as its summary (#42)
  • interface: a PWA on the phone, and the rail and drawer become sheets (#20)

Fixes

  • chats: end the first turn at done, and name the chat after it (#226) (#228)
  • web: name Hera, and say she only after naming her (#223) (#227)
  • chat: make Stop end the turn and give the composer back (#216) (#219)
  • ci: /preview\ should wait for the capture, not race it (#139)
  • web: correct 'add model' button text (#123)
  • web,core: artifact drawer state, raw HTML in prose, and the preferred model list (#97)
  • providers: read reasoning from whichever field the server sends (#96)
  • web: manage servers button opens the servers tab, not models (#91)
  • providers,web: stop announcing a whitespace-only thought (#90)
  • rebuild each round's messages from the record instead of accumulating them (#67)
  • interface: disambiguate the composer's two context pills for e2e (#37)
  • web: bump vitest to ^4.1.11 for vite 8 compatibility (#36)
  • interface: detach hera__ask from the emotion vocabulary, then drop it (#21)

Documentation

  • point the Docker example at :rc until the first release exists (#213)
  • rework the README and refresh CONTRIBUTING (#200) (#202)
  • add Publii website pages, drop drafts file (#39)
  • cut status.md prose, translate spec docs to English, stage Publii drafts (#27)

Maintenance

Other

  • Update README.md (#178)
  • Show only configured MCP servers in the composer (#26)

Full changelog: v0.2.0...v0.2.1