Download Latest Version v1.2.75 source code.tar.gz (455.4 kB)
Email in envelope

Get an email when there's a new version of The Pope Bot

Home / v1.2.74
Name Modified Size InfoDownloads / Week
Parent folder
README.md 2026-03-30 8.7 kB
0
v1.2.74 source code.tar.gz 2026-03-30 504.0 kB
0
v1.2.74 source code.zip 2026-03-30 760.1 kB
0
Totals: 3 Items   1.3 MB 0
  • fix: auto-replace unsupported Gemini thinking models with gemini-2.5-flash (#202)
  • feat: settings overhaul, DB-backed config, dynamic code tabs
  • fix: merge [#202] Gemini thinking model fallback, update provider list
  • feat: reorganize settings page layout, combine LLM tabs
  • feat: restructure docker, admin routes, API keys voice tab, auth roles
  • fix: use chrome instead of chromium for Playwright MCP, remove stale .env version sync
  • fix: use chromium with --with-deps for Playwright MCP in claude-code-job
  • fix: chown pw-browsers dir, simplify setGitHubSecret
  • fix: restore secret encryption, add Secrets permission, sync GH_TOKEN to .env
  • feat: add delete buttons for GitHub secrets and variables
  • feat: bake .next into Docker image, remove app/ from user projects
  • fix: use minimal package.json in Docker build/runtime stages
  • fix: resolve thepopebot self-install and module resolution in Docker build
  • fix: force-recreate event-handler container on upgrade
  • feat: show real isSet status for GitHub secrets and variables
  • fixes to system
  • fix: add safe.directory for /project in workflows
  • fix: rebuild workflow git config and compose paths
  • fix: add --no-install flag to init, use in rebuild workflow
  • fix: remove --no-install flag, always run npm install during init
  • fix: restore --no-install flag, add set -e to workflow exec blocks
  • fix: move safe.directory config before gh auth in workflows
  • refactor: unify settings UI with shared components and design standards
  • fix: remove set -e from workflow exec blocks
  • fix: use /project paths for docker compose in rebuild workflow
  • fix: respect COMPOSE_FILE env var in rebuild workflow
  • feat: add toolbar to secondary code/shell tabs, plan/code mode toggle, and headless improvements
  • feat: move code mode toggle to runtime state and clean up headless streaming
  • refactor: rewrite headless stream parser to use split2 and Transform
  • fix: preserve code workspace volumes across container lifecycle
  • chore: remove unused CODE_EXECUTION.md template and path
  • fix: add newline separator between tool response and headless stream
  • feat: improve Working spinner and tool call spacing in chat UI
  • feat: add configurable debounce for cluster file watch trigger
  • feat: add ConvertKit email signup, investigate tool, cluster plan mode, and Docker networking
  • fix: prefix investigate container name with claude-code-
  • fix: use correct ConvertKit form ID for email signup
  • added docs for scafolded projects
  • revert: remove hardcoded Docker network, restore auto-detection
  • refactor: overhaul chat agent tools — fix headless streaming, rename tools, remove dead code
  • plan for docker upgrade
  • fix: pass missing slug prop to ProviderCard on ChatLlmPage
  • fix: improve provider spacing and consolidate OAuth token storage
  • Add browser-based file editor tab to code workspaces
  • feat: improve browser editor with theme sync, stale detection, and inline UI
  • feat: unified coding agent Docker image with multi-agent support
  • refactor: add token type param to OAuth functions and use dialog for token creation
  • feat: coding agents settings page with multi-agent container integration
  • feat: add Gemini CLI, Codex CLI, and OpenCode coding agents (headless runtime)
  • feat: red collapsible box for unknown headless stream events
  • feat: add DeepSeek, MiniMax, Mistral, xAI, OpenRouter providers and update model lists
  • feat: pretty-print stage names in coding agent entrypoint
  • feat: third-party backends for Claude Code, containers page, and admin restructure
  • feat: rename workspace runtime to interactive, simplify clone and close flows
  • feat: workspace command runtimes, diff stats, and code mode persistence
  • fix: use gap spacing instead of text space in diff stats badge
  • feat: replace Docker volumes with bind-mount dirs, rename container functions, and improve streaming
  • feat: diff viewer, workspace command overhaul, and unified tool prompt field
  • fix: diff stats compare against feature branch, refresh before commands
  • feat: base branch selector, hasChanges tracking, and entrypoint stdout fix
  • fix: diff stats fallback to base branch, fix untracked file line counting, remove get_repository_details tool
  • fix: improve code workspace change indicator visibility
  • feat: code workspace toolbar overhaul, diff viewer, voice input, and singleton fixes
  • refactor: rename job→agent-job and make all chats workspace chats with singleton agents
  • fix: convert all mount/interval server actions to fetch routes to prevent page refresh
  • fix: convert getRepositories server action to fetch route and fix code mode repo state
  • feat: agent job secrets UI, unified Docker images, and fetch route migration
  • feat: port forwarding for code workspaces via Traefik
  • fix: port forward cleanup and duplicate prevention
  • feat: Kimi provider, Codex auto-config, sync overhaul, and workspace cleanup
  • feat: add chat mode context (plan/code/job) to agent chat
  • fix: missing chat data route, rename workspace commands, sidebar icon
  • fix to local storage and setting recent github command + gemini cli settings
  • feat: OAuth helper for agent job secrets
  • fix: OAuth helper cleanup and chat mode restoration
  • fix: use stable random feature branch names to prevent race condition
  • refactor: rename tools to agent_job and coding_agent
  • updates gitignore
  • fix: align Docker image names with product names and parallelize builds
  • feat: unify add secret dialog with Manual/OAuth toggle and rename route to agent-jobs
  • feat: replace provider select with searchable combobox
  • feat: inject agent job secrets into headless and interactive containers
  • fix: chown workspace dirs so coding-agent container user can write
  • fix: capture agent stdout/stderr to log files in agent-job runtime
  • fix: also create and chown workspace/ subdir before container start
  • fix: remove stdout/stderr fd redirection from coding-agent entrypoint
  • feat: session continuity across headless/interactive + secrets & skills refactor
  • fix: disable tmux mouse mode to restore terminal scrollback, add WebSearch/WebFetch permissions, inline voice token fetch
  • feat: track Claude Code session IDs via .claude-ttyd-sessions/ for tab reconnection
  • fix: use port-based labels for terminal tabs to prevent duplicates
  • refactor: rename extra-tab.sh to start-code-session.sh, extract start-shell-session.sh to common
  • fix: remove stray brace left from typeCounts removal in listTerminalSessions
  • fix: verify session jsonl exists before resuming, fall back to fresh session
  • fix: always overwrite session ID in port file on SessionStart
  • refactor: split ttyd startup from per-connection session logic for dynamic resume
  • feat: wrap code workspace tabs in tmux to survive browser disconnects
  • fix: pass LLM_MODEL to all coding agent interactive sessions
  • docs: update CLAUDE.md files to reflect current architecture
  • fixes to tmux
  • added tmux
  • fix: pass PORT env var into tmux session in start-coding-session.sh
  • feat: add manual check for updates button and remove keyboard tab sorting
  • feat: persist active tab and select left neighbor on close in code terminal
  • fix: add OSC 52 clipboard support and terminal resize on reconnect
  • Revert "feat: persist active tab and select left neighbor on close in code terminal"
  • fix: use sendResize for tab activation and add tmux copy-mode bindings
  • chore: default GitHub runner replicas to 1
  • feat: consolidate Docker log streaming with shared parser and SSE endpoint
  • fix: include prompt in tool return values for agent and coding tools
  • fix: comprehensive mobile UI fixes and sidebar dropdown portal
  • removed test file
  • chore: remove audit screenshots and gitignore screenshots dir
  • fix: show complete tool call args in chat UI by accumulating streaming chunks
  • feat: enable all providers as Claude Code backends via LiteLLM proxy
  • fix: add Docker URL help text to custom provider Base URL field
  • feat: add Google/Gemini as LiteLLM proxy backend for Claude Code
  • fix: make docker-compose.litellm.yml a full working compose file
  • fix: load .env into process.env during setup and remove build step
  • fix: move default database path to data/db/thepopebot.sqlite
  • feat: unify custom provider model selector and clean up litellm compose
  • feat: support multiple models per custom provider
  • feat: text input with save button for providers with no model list
  • feat: unified save feedback and text-input Save button for LLM settings
  • refactor: use card-level save indicator for all save actions
  • fix: unify LLM missing-key errors into single generic message
  • docs: comprehensive documentation update for current architecture
Source: README.md, updated 2026-03-30