This Week in gptme (W24 2026)

Here’s what landed in gptme and gptme-contrib this week (2026-06-08 – 2026-06-14): 76 new features, 43 bug fixes across 145 merged PRs.

Highlights

gptme#2885 (commands) /snapshot — agents can now checkpoint their conversation tree and resume from any past state
gptme#2882 (tools) view_anchored and patch_anchored — edit files without line-number drift
gptme#2878 (commands) /backtrack — in-session conversation recovery; roll back to any prior checkpoint
gptme#2880 (tools) ToolFunction abstraction — structured callable metadata for tools, enabling richer composition
gptme#2778 (webui) browser TTS for assistant messages via Web Speech API

New Features

gptme#2778 (webui) add browser TTS for assistant messages (Slice 1 — Web Speech API)
gptme#2779 (webui) add Alt+N global shortcut for new conversation
gptme#2781 (plugins) discover src-layout plugins, layer user [plugins], lenient +tool
gptme#2791 (server) add detail query param to conversation list endpoint
gptme#2792 (webui) auto-poll provider health every 60s; add degraded badge on settings icon
gptme#2794 (webui) add OpenRouter STT fallback for unsupported browsers
gptme#2795 (webui) add server-side TTS endpoint via OpenRouter
gptme#2797 (cli) add –format table option to gptme-util status
gptme#2801 (webui) model favorites, set-as-default, and picker scroll fix
gptme#2803 (webui) toggle/stop TTS playback from the message button
gptme#2806 (webui) add /settings route with full-page settings view
gptme#2812 (webui) show active model in conversation header
gptme#2813 (ci) self-heal Phase 2 — auto-open draft fix PR on structured gate pass
gptme#2814 (artifacts) include workspace files created/modified by the conversation
gptme#2815 (webui) add TTS engine selector (browser / gptme-server / gptme-tts)
gptme#2816 (webui) add keyboard shortcut (Ctrl+Shift+) to toggle split view
gptme#2817 (webui) add per-pane conversation selector in split view (Slice 2)
gptme#2818 (artifacts) format-independent file-write parsing + diffs
gptme#2823 (webui) show message count as badge on conversation title line
gptme#2824 (webui) add URL subroutes for settings category deep-linking
gptme#2825 (webui) add persistent no-confirm (YOLO) mode setting
gptme#2827 (webui) add conversation list search/filter
gptme#2828 (webui) show relative timestamps on messages with auto-refresh
gptme#2831 (config) add context_exclude glob patterns to [prompt] in gptme.toml
gptme#2832 (webui) highlight matched search terms in conversation list titles
gptme#2833 (server) expose message_count and last_updated in conversation list
gptme#2834 (webui) persist conversation search query in URL ?search= param
gptme#2837 (eval) add narrow swebench install extra
gptme#2838 (webui) add / keyboard shortcut to focus conversation search
gptme#2840 (webui) add settings-level STT provider choice (browser vs server)
gptme#2841 (webui) show conversation stats in list using list-response data
gptme#2843 (cli) add startup prompt stats report
gptme#2844 (eval) surface per-task tool-call count (tool-efficiency metric)
gptme#2848 (webui) server-backed star/unstar for conversations via metadata sidecar
gptme#2853 (cli) add –no-workspace flag to skip all workspace context
gptme#2860 (server) cursor-based pagination for conversations list
gptme#2861 (server) add limit/before message pagination to conversation GET
gptme#2864 (server) add ?q= filter param to conversations API with server-side webui search
gptme#2865 (webui) add sort control to conversation list (Recent/Longest/A-Z)
gptme#2869 (openai) enable Responses API by default for gpt-5/o-series models
gptme#2870 (server) add /api/v0/metrics Prometheus endpoint
gptme#2872 (webui) thread log_offset through conversation state for windowed pagination (Slice 1)
gptme#2878 (commands) add /backtrack for in-session conversation recovery
gptme#2880 (tools) introduce ToolFunction abstraction for structured callable metadata
gptme#2881 (subagent) forward subprocess/isolated defaults from planner subtask roles
gptme#2882 (tools) add view_anchored and patch_anchored tools (Gate G2)
gptme#2884 (computer) add scroll action for mouse wheel support
gptme#2885 (commands) add /snapshot command for agent tree search (#495)
gptme#2886 (snapshot) embed conversation message count for diff summary
gptme-contrib#1061 (gptme-codegraph) add local-first semantic search (Phase 1)
gptme-contrib#1063 (gptme-coordination) add generic inter-agent coordination package
gptme-contrib#1065 (gptme-tts) actionable available_hint when TTS is unavailable
gptme-contrib#1069 (trimmer) add tool-output summarization pass (priority 199)
gptme-contrib#1070 (protocols) add GPTME_HEARTBEAT spec, schema, and validator
gptme-contrib#1074 (imagen) register image_gen as a gptme.plugins entry point
gptme-contrib#1075 (pm-dispatch) per-lane model routing (fast lane → cheaper model)
gptme-contrib#1077 (pm-dispatch) Thompson-sampling bandit for per-lane model routing
gptme-contrib#1078 (fleet-vitals) shared emit helper for the fleet-vitals contract
gptme-contrib#1079 (imagen) add execute function for image_gen blocks with inline webui preview
gptme-contrib#1080 (self-merge) Greptile score floor — upstream greptile-merge-signal + wire into gate
gptme-contrib#1081 (pm) surface own-PR Greptile review via notification-blind detector
gptme-contrib#1082 (github) add resolve-greptile-threads.py (resolveReviewThread primitive)
gptme-contrib#1083 (gptodo) add ‘status –json’ for machine-readable output
gptme-contrib#1085 (agent-msg) add reply command and needs-reply tracking
gptme-contrib#1086 (scripts) upstream codex + openrouter usage scrapers
gptme-contrib#1088 (subscription) upstream config-driven check-quota.py + shared harness_models
gptme-contrib#1089 (gptmail) add channel field to MessageInfo tracker
gptme-contrib#1090 (gptmail) add Transport Protocol + EmailTransport adapter (fold step 1)
gptme-contrib#1092 (self-merge) capture headRefOid in CheckResult for pre-merge re-check
gptme-contrib#1094 (gptmail) add AgentTransport (fold step 2)
gptme-contrib#1097 (gptmail) add gptmail agent CLI subgroup (fold step 3)
gptme-contrib#1098 (subscription) add HarnessQuotaConfig + load_quota_config() for agent-generic quota
gptme-contrib#1101 (usage) new gptme-usage package — move harness_models + config-leak fixes
gptme-contrib#1102 (usage) config-driven quota — ship generic, data via per-agent harness-quota.toml
gptme-contrib#1103 (gptmail) unread-default + read markers for gptmail agent list
gptme-contrib#1108 (gptme-usage) add merge_with_module_defaults helper

Bug Fixes

gptme#2777 (ci) pin self-heal logs to triggering attempt
gptme#2793 (plugins) recurse nested src-layout tool packages
gptme#2796 (webui) request fast conversation list detail
gptme#2798 (cli) reject malformed model paths
gptme#2800 conversation load crash chain (server symlink 500 + webui dataless crash + error surfacing)
gptme#2802 (webui) UI polish (input radius, code blocks, avatar, health badge)
gptme#2804 (webui) sync model selector when switching conversations
gptme#2807 (hooks) hide injected agent-instructions from conversation view
gptme#2808 (webui) don’t flush queued messages during tool confirmation
gptme#2809 (tools) report unavailable tools accurately instead of ‘invalid choice’
gptme#2820 (cloud) route LLM API calls to Supabase edge functions, not fleet.gptme.ai
gptme#2821 (server) graceful shutdown on SIGTERM
gptme#2822 (webui) make Escape key and interrupt logs consistent with isBusy
gptme#2826 (cloud) parse max_completion_tokens for gptme provider models
gptme#2829 (webui) make clickable selection rows keyboard-accessible
gptme#2835 (webui) update ConversationList test to handle highlighted search terms
gptme#2845 (webui) add Escape-to-navigate-back for full-page settings view
gptme#2846 (complete) auto-reply kill-switch now detects incomplete-todos variant
gptme#2849 (webui) add semantic landmarks, aria-current, and aria-selected for keyboard navigation
gptme#2856 (server) scope conversations cache to active logs dir
gptme#2857 (webui) reduce redundant conversations API calls on page load
gptme#2859 (server) preserve messages field on cached conversation list responses
gptme#2867 two startup hangs — skip OpenRouter fetch for bare model names, non-blocking piped stdin
gptme#2868 (server) validate model type in config PATCH to prevent 500
gptme#2871 (models) validate provider default model entries + update groq default
gptme#2873 (server) improve model validation error messages with examples
gptme#2874 (server) tolerate pre-existing workspace symlink in ChatConfig.from_logdir
gptme#2877 (server) return 400 for non-positive limit in conversations API
gptme#2879 (webui) collapse single hook system messages in chat view
gptme-contrib#1062 (gptme-tts) make plugin discoverable and fix stale hook names
gptme-contrib#1064 (subscription) export pacing snapshot helpers
gptme-contrib#1066 (self-merge-check) accept ‘owner/repo ' two-positional form
gptme-contrib#1071 (pr-queue-health) treat SKIPPED/NEUTRAL checks as passing
gptme-contrib#1072 (gptodo) allow editing tracking_issue and upstream_coordination_id via –set
gptme-contrib#1084 (scripts) reconcile check-claude-usage.sh with main-repo fixes
gptme-contrib#1093 (lessons) remove dead keywords from project-monitoring-session-patterns
gptme-contrib#1095 (scripts) treat partial-null limit_remaining as available in check-openrouter-usage
gptme-contrib#1096 (monitoring) emit own-PR Greptile review for BLOCKED PRs
gptme-contrib#1104 (self-merge) gate eligibility on actual merge permission
gptme-contrib#1106 (runloops) lazy pm_dispatch re-export to silence runpy RuntimeWarning
gptme-contrib#1107 (self-merge) scope spec-like-doc gate to repo root only
gptme-contrib#1109 (gptmail agent) report delivery failures + guard self-send
gptme-contrib#1110 (gptmail agent) exclude peer replies to my messages from pending

Performance

gptme#2852 (server) enable gzip compression on API responses
gptme#2854 (server) skip branch glob on fast scan, dedupe stat, drop debug log
gptme#2862 (tools) lazy-load mcp.types and bashlex to reduce startup time
gptme#2863 (server) add ETag conditional request support for conversations endpoints
gptme#2866 (tools) lazy-load MCPClient and MCPRegistry to reduce startup time

Refactors

gptme#2805 (webui) group Settings → Servers into collapsible sections
gptme#2875 (llm) extract shared Responses API event loop into openai_responses.py
gptme-contrib#1105 (agent-msg) convert to thin shim over gptmail agent CLI

Documentation

gptme#2782 (howto) add how-to guide section with 5 task-oriented recipes
gptme#2810 add model-routing guide — pick your model per task
gptme#2811 (config) document [models] section — default and favorites keys
gptme#2858 (howto) remove inaccurate –no-workspace equivalence claim

Tests

gptme#2780 (tauri) add E2E chat interaction flow tests (Slice 2)
gptme#2830 (server) add PATCH edit-message and DELETE message edge-case tests
gptme#2839 (webui) add Playwright e2e keyboard navigation tests
gptme#2847 (complete) cover cross-variant auto-reply exit counting
gptme-contrib#1068 (coordination) add tests for db module
gptme-contrib#1073 (gptme-runloops) add tests for utils/prompt utilities

CI & Infrastructure

gptme#2784 (deps) bump actions/cache from 4 to 5
gptme#2785 (deps) bump codecov/codecov-action from 6 to 7
gptme#2786 (deps-dev) bump pre-commit from 4.0.1 to 4.6.0
gptme#2787 (deps-dev) bump mypy from 1.13.0 to 2.1.0
gptme#2788 (deps) bump datasets from 4.6.0 to 5.0.0
gptme#2789 (deps-dev) bump sphinx-book-theme from 1.1.3 to 1.1.4
gptme#2790 (deps-dev) bump myst-parser from 4.0.0 to 4.0.1

Chore

gptme-contrib#1067 ignore Python build/ and dist/ artifacts (git + jscpd)

*145 PRs merged across 2 repos. See the full changelogs: gptme

gptme-contrib*