CharacterVault 1.8.0
CharacterVault 1.8.0 makes the AI Agent easier to steer, follow, and review: edit and resend your last message, @-mention fields and entries, see what each run is doing and what earlier runs changed, and check its edits in a clearer diff.
Try it: https://vault.charactervault.appStaging: https://staging.charactervault.appDocs: https://vault.charactervault.app/docs/
Highlights
Steer the Agent from the chat
Redirect the Agent without retyping or starting over. Everything here lives in the Agent chat on a character or lorebook.
- Edit and resend — the pencil on your last message puts its text back in the box. Nothing is removed until you send; then that message and the replies after it are replaced. The × cancels and keeps the chat as it was.
- @-mentions — type
@in the composer to pick a field, alternate greeting, or lorebook entry. It inserts plain text such as@Description,@Greeting 2, or@“Harbor” (#4), and the Agent reads it as that item. ↑ / ↓ move, Enter or Tab picks, Esc closes the list. - Starter suggestions — an empty chat offers first requests based on what the card or book is missing, such as Write a description or Build a lorebook from my custom context. It never suggests Appearance, Personality, Scenario, System prompt, or Post-history.
- Continue — a run that stops at the 32-turn limit says so on its last message, with a Continue button that picks up where it left off.
Know what the Agent is doing
The Agent now tells you what it is working on and remembers what it already did.
- Status line — while a tool runs, the spinner names the target, such as Updating Description, Reading entry “Harbor”, or Searching “harbor”.
- Run memory — each new request carries a short note on what earlier runs in the chat changed and whether you applied or discarded their review, so “undo that” or “do the same for greeting 2” work. New chat starts over.
- Finished in the background — if a run ends while you are on another tab or in another app, the page title starts with Agent finished until you come back.
AI Agent → What you see in chat
Clearer review diffs
Turn on Settings → Character Workspace → Review agent edits before applying (off by default) to approve changes before they land.
- Side by side or inline — wider screens show Original and Agent next to each other, aligned line by line; phones get one inline column.
- Word highlights — reworded lines highlight only the words that changed.
- Folded context — runs of unchanged lines collapse behind Show N unchanged lines.
- Header status — a shield icon shows review is on, and a yellow Review N button reopens a waiting proposal.
Also in this release
- Rate limits, server errors, and dropped connections retry automatically, up to 4 times with backoff, and honor Retry-After.
- Long runs trim older tool results to fit Settings → Sampler → Context Length instead of overflowing.
- Snippet replaces fall back to a case-insensitive match, and truncated or misnamed native tool calls are handled more gracefully.
- New chat is an icon, and the Native / XML chip hides when the chat panel is narrow.
- The Agent docs now list the lorebook entry flags and book settings it can write.
Fixes
- Edits you make in the editor while the Agent runs are no longer overwritten when it writes. Review Apply also lands on the latest card.
- Agent updates that change nothing no longer count as writes or trigger a save.
Upgrade
Client-only app. Open the hosted site or rebuild locally; no server migration.
git clone https://github.com/spaceman2408/CharacterVault
cd CharacterVault
npm install
npm run devLocal data from 1.7.5 keeps working. Hard refresh if the UI looks stale after deploy.