Releases: wesm/agentsview
v0.23.1
v0.23.0
-
Add live dashboard refresh via Server-Sent Events so sessions and usage update without reloading.
-
Add session intelligence with health signals, analytics, and richer UI insights.
-
Add a
sessionCLI for listing, inspecting, exporting, syncing, and watching session data programmatically. -
Add Markdown export for sessions.
-
Add SSH remote session sync to pull in sessions from remote machines.
-
Add PostgreSQL support for usage reporting.
-
Add
agentsview statsto report per-author git and session activity. -
Require the explicit
servesubcommand to start the server. -
Show session names in the Top Sessions by Cost view.
-
Preserve active filters when switching between the Sessions and Usage tabs.
-
Link usage group-by selectors for a smoother analytics workflow.
-
Make incremental parsing more reliable by validating files with inode and device tracking.
-
Rewrite the README as a clearer user-facing guide.
-
Surface the underlying login shell probe failure reason for the desktop app on arm64 flake cases.
-
Only self-heal the events store on transient frontend failures.
-
Accept raw session IDs in
token-use. -
Resolve ephemeral serve port handling correctly.
-
Parse Claude Code CLI JSON array output correctly in insights.
-
Recognize the
"helpful assistant"prefix as an automated session. -
Skip git-root walking for foreign-OS working directories.
v0.22.2
- Correct fallback pricing for Claude Opus 4.6 so usage cost estimates use the proper
$5input and$25output rates.
v0.22.1
- Keep Codex model detection consistent during incremental syncs so usage reporting stays accurate.
v0.22.0
-
Track usage costs for OpenCode and Pi in summaries and cost charts.
-
Improve usage dashboard cards and time-series charts for clearer cost reporting.
-
Sync Codex exec sessions so execution activity appears reliably in agentsview.
-
Fix usage accounting so total cost data stays complete across supported providers.
v0.21.0
-
Add a token usage dashboard with filtering and drill-down views for exploring usage by project, model, and session.
-
Show usage summaries, cost trends, attribution, cache efficiency, top sessions, and treemap breakdowns in the new dashboard.
-
Add a dedicated usage view and API support for browsing token and cost data more easily.
-
Prevent some Claude usage records from being double-counted in totals.
-
Correct Opus 4.6 pricing in usage cost calculations.
v0.20.0
-
Add support for OpenHands CLI sessions, with a shallow watch mode for lighter syncing.
-
Add support for Positron Assistant sessions.
-
Show estimated usage cost in
agentsview usage dailyand the status line. -
Filter pinned messages by the selected project.
-
Make
agentsview usage dailyfaster and default it to the last 30 days. -
Simplify docs and help text to use broader agent guidance instead of long agent lists.
-
Restore Intel macOS desktop builds in release artifacts.
-
Recover the desktop app automatically if the WebView dies after sleep/wake.
-
Fix configuration key mapping for Pi, Cursor, and Amp agents.
v0.19.0
-
Add support for Warp agent sessions.
-
Add support for Hermes Agent sessions.
-
Add support for Cortex Code sessions.
-
Add support for Kiro CLI and Kiro IDE sessions.
-
Add Cursor session history with resume and workspace recovery.
-
Add the ability to resume Claude sessions in Claude Desktop.
-
Add project filtering to
pg push. -
Add an opt-out for update checks and expand privacy and telemetry documentation.
-
Fix the synced status label staying stale and show the exact sync time on hover.
v0.18.0
-
Import conversation archives from Claude.ai and ChatGPT.
-
Track token usage across sessions, analytics views, and exports.
-
Redesign the import modal and stream live progress while imports run.
-
Detect automated RoboRev sessions and filter them out of normal browsing and analytics views.
-
Preserve pinned messages after a session resync.
-
Reduce desktop WebView memory usage when rendering session content.
v0.17.1
Bug Fixes
pg serveautomatically applies pending PostgreSQL schema migrations on startup, reducing upgrade friction and startup failures.