openagentd 1.99.3

Self-hosted AI agents — streaming chat, tool use, persistent memory, multi-agent teams

OpenAgentd

The desktop cockpit for local AI agents. A double-click app that runs a team of AI agents on your machine, with a real UI to watch every step. Multiple windows, coding workspaces, git worktrees, scheduled tasks, local telemetry, and your own keys. Open source.

Features · Issues / roadmap · Documentation · Migration

---

Screenshots

Home	Split team cockpit

Coding workspace	Diff inspector

Providers settings	Sandbox settings

Scheduler	Telemetry dashboard

Command palette	Shell mode

Git commit tree	Terminal tab

Session settings	Image lightbox

Todos board	MCP settings

Mobile backend settings	Mobile coding view

---

What you get

A short list of the most important shipped capabilities. The canonical, version-cited catalogue is documents/docs/features.md; other feature requests and future work live in GitHub issues.

A cockpit, not a chat box. OpenAgentd gives you the missing UI around tool-using agents: command palette (⌘P/Ctrl+P), slash commands (/init, /continue, /compact, /undo, /redo), direct shell sends with !command, drag-and-drop attachments, a shared lightbox for generated and uploaded media, and a tool inspector that exposes arguments, status, timing, results, and inline Git-like diffs.

Run a team, not just one agent. A lead agent can spawn specialist instances on demand, coordinate them through an async mailbox, and manage what tools, skills, or MCP servers each member can use. You can watch lead and worker sessions side by side in the split cockpit, follow each agent's output live, collapse back to a unified chronological transcript, or open multiple desktop windows when one conversation is not enough.

Use it as a coding cockpit. Coding mode opens any local project folder with a workspace-aware team, a repo tree, live git changes, commit history/tree views, file previews, and an interactive terminal tab in one place. Create isolated git worktree sessions from the UI, inspect staged/unstaged diffs in the right-side dock, browse recent commits and branch structure, open @file mentions directly into the preview panel, and jump into shell mode without leaving the composer. After an agent edits code, OpenAgentd runs the project's own language servers over the change and feeds diagnostics back into the tool result so the next turn can fix them immediately.

Pick your model, no lock-in. OpenAgentd ships with 15 first-class providers — Anthropic, Gemini, OpenAI, OpenRouter, Bedrock, Grok, DeepSeek, Ollama, and more — while keeping the config format simple: provider:model. Manage provider connections from Settings, choose which models stay visible in pickers, and override the lead model or thinking level per session from Session Settings. Assistant replies keep the effective model in history so overrides stay auditable.

Desktop + mobile clients. The main experience is the desktop cockpit on macOS and Linux, and there is also a Tauri mobile app that connects to a remote OpenAgentd server over LAN or the internet. Mobile uses the same shared UI, adapted for touch, safe areas, backend switching, and single-pane navigation, including coding sessions on a phone-sized screen.

Operate locally, extend when needed. Scheduled tasks support cron, interval, and one-shot runs; todos stay on a live board; telemetry stays on your machine; MCP servers hot-reload from Settings; and sandbox rules plus permission decisions are managed without editing hidden app internals by hand.

---

Why OpenAgentd

Coding agents (Claude Code, Codex CLI, Cursor, Windsurf, Aider, opencode) all run agents with tools now. The real difference is the shape of the workflow they fit into — terminal session, IDE window, or a desktop cockpit.

	openagentd	Claude Code	Codex CLI	Cursor / Windsurf
Surface	Desktop app + mobile app + web cockpit	Terminal (CLI)	Terminal (CLI)	IDE (VS Code fork)
Multi-agent	Lead + workers, split-pane live view	Sub-agents (sequential)	—	—
Watch live	Tool inspector + diffs + per-call timing	Terminal text	Terminal text	Inline in editor
Direct shell from input	!command shell mode + structured history	CLI shell escapes	CLI shell escapes	Terminal/editor tasks
Git worktree sessions	Managed worktrees as repo children in UI	Manual git setup	Manual git setup	IDE/git extension flow
/undo across chat	Restores workspace files from any prior turn	—	—	Editor undo only
Providers	15 — bring your own keys	Anthropic only	OpenAI only	A few, subscription
License / cost	Apache 2.0, your keys	Proprietary + sub	Proprietary + sub	$20/mo+ subscription

Full capability matrix (incl. Aider + opencode): documents/docs/comparison.md. Canonical feature catalogue with version-cited ship dates: documents/docs/features.md.

---

Quick start

Desktop app (double-click install, no terminal) — download from the latest release. OpenAgentd is the primary desktop cockpit on macOS and Linux, with support for multiple app windows and coding worktree sessions:

Platform	Artefact	First-launch note
macOS (Apple Silicon)	brew install --cask lthoangg/tap/openagentd or OpenAgentd_*_aarch64.dmg	The cask ad-hoc signs and installs automatically. With the .dmg, run the bundled install.sh then right-click → Open. The app is unsigned — why.
Linux	OpenAgentd_*_amd64.AppImage or OpenAgentd_*_amd64.deb	Run the AppImage directly (chmod +x, then execute it) or install the .deb with your package manager, e.g. sudo apt install ./OpenAgentd_*_amd64.deb.

macOS — after mounting the .dmg:

cd /Volumes/OpenAgentd*
./install.sh --install            # ad-hoc signs + copies to /Applications

Then right-click OpenAgentd.app → Open the first time (single-click won't work).

CLI / API server (terminal install — backend only; connect from the desktop app via Server connection):

# macOS / Linux
uv tool install openagentd        # recommended
brew install lthoangg/tap/openagentd
curl -fsSL https://raw.githubusercontent.com/lthoangg/openagentd/main/install.sh | sh   # zero-setup: bootstraps uv, then installs

openagentd init   # pick provider + API key, install default agents
openagentd        # API server on http://localhost:4082

For phones or another desktop on the same network:

openagentd start --lan --key   # bind 0.0.0.0, save an access key, and print the LAN/mobile URL
openagentd address       # show local + LAN URLs again later
openagentd health        # verify the backend is reachable and ready

Mobile app — OpenAgentd also ships a Tauri mobile client for connecting to a remote OpenAgentd server. Use openagentd start --lan --key on your host machine, then connect from the phone with the same backend-connection flow. See documents/docs/mobile.md.

Other install options (pip, pipx, from source) — see documents/docs/install.md.

Useful maintenance commands:

openagentd start --lan --key    # expose backend to desktop/mobile on your LAN with access-key auth
openagentd restart              # restart the background server
openagentd address              # show local and LAN server URLs
openagentd health               # run server/mobile diagnostics
openagentd status               # show PID, URLs, and log path
openagentd logs                 # tail the local server log
openagentd doctor               # check install health
openagentd cleanup              # dry-run cleanup for generated artifacts older than 14 days
openagentd cleanup --apply      # delete the listed generated artifacts
openagentd upgrade              # stop, upgrade, and restart if running

Generated artifacts are session-scoped under {OPENAGENTD_DATA_DIR}/sessions/{session_id}/, not inside the active workspace or coding repo. Todos live at .todos.json inside that session artifact directory; bulky tool output lives under .tool_results/, including shell spills at .tool_results/shell/. Deleting a normal session removes its workspace and session artifact directory. Deleting a coding session keeps the project directory and removes only that session's app-managed artifacts.

---

Migrate from OpenClaw or Hermes Agent

Import existing identity and context Markdown files into one OpenAgentd lead agent:

openagentd migrate openclaw --from ~/.openclaw/workspace --model openai:gpt-5.5
openagentd migrate hermes --from ~/.hermes --model openai:gpt-5.5

Existing agent files are not overwritten unless --force is passed. See MIGRATION.md for setup migration from OpenClaw, Hermes, Claude Code, and Codex CLI.

---

Providers

Switch models with a single line in your agent's .md config file. Every provider uses the provider:model format, and Settings → Providers can manage OAuth connections, API-key-backed providers, visible-model filters, and per-provider status from the UI.

Provider	Format	Auth
Anthropic Claude	anthropic:claude-sonnet-4-6	ANTHROPIC_API_KEY
Google Gemini	googlegenai:gemini-3.1-flash	GOOGLE_API_KEY
Google Vertex AI	vertexai:gemini-3-flash-preview	VERTEXAI_API_KEY or GCP creds
OpenAI	openai:gpt-5.5	OPENAI_API_KEY
OpenRouter	openrouter:qwen/qwen3.6-plus:free	OPENROUTER_API_KEY
ZAI / GLM	zai:glm-5-turbo	ZAI_API_KEY
xAI Grok	xai:grok-4.20	XAI_API_KEY
DeepSeek	deepseek:deepseek-v4-flash	DEEPSEEK_API_KEY
AWS Bedrock	bedrock:anthropic.claude-sonnet-4-6	AWS profile / default chain
NVIDIA NIM	nvidia:stepfun-ai/step-3.5-flash	NVIDIA_API_KEY
GitHub Copilot	copilot:gpt-5.4-mini	openagentd auth copilot
OpenAI Codex	codex:gpt-5.5	openagentd auth codex
Router9 (local)	router9:cc/claude-sonnet-4-5	ROUTER9_API_KEY (optional ROUTER9_BASE_URL)
CLIProxyAPI (local)	cliproxy:gemini-2.5-pro	CLIPROXY_API_KEY (optional CLIPROXY_BASE_URL)
Ollama (local + cloud)	ollama:llama3.2 · ollama:kimi-k2.6-cloud	none (cloud: ollama signin)

In the cockpit, Session Settings can override the lead agent's model and thinking level for the current chat; history keeps the model used for each user turn, and assistant reply footers show the effective model that produced the response.

---

Built-in tools

OpenAgentd starts with a compact local toolbelt for files, shell, web, generation, scheduling, and team coordination:

Category	Tools
Filesystem	read, write, edit, patch, ls, glob, grep, rm
Shell	shell, bg (background processes)
Web	web_search, web_fetch
Generation	generate_image, generate_video
Scheduling	schedule_task
Tasks	todo_manage
Team coordination	team_message, team_manage (teams only)
Utility	date, skill

Add any MCP server to expose more tools without writing custom app code.

---

Agents and teams

OpenAgentd ships with a compact cockpit team. The desktop app can open multiple windows, while coding sessions can branch into separate git worktree-backed workspaces when you need isolated changes:

Agent	Role	Specialty
openagentd	Lead	Coordinates the team, receives user messages, spawns members, delegates
executor	Member blueprint	File creation, builds, shell commands, tangible artifacts
explorer	Member blueprint	Web research, codebase exploration, information gathering

Configure your team by editing .md files in your config directory. Exactly one agent must have role: lead; the rest are member blueprints. The lead uses team_manage to spawn/dismiss live instances (executor#1, explorer#1) and team_message to delegate and collect results.

Fresh installs also seed a separate coding team under agents/coding/. Open /coding to select a server-local project folder and start workspace-aware sessions; Settings shows those agents as coding/openagentd, coding/coder, and coding/explorer.

Agent config at a glance

---
name: my-agent
role: member
description: Handles deep research tasks
model: googlegenai:gemini-3.1-flash
thinking_level: high
tools:
  - web_search
  - web_fetch
  - read
  - note
mcp:
  - context7
---

System prompt goes here.

Member instances are created lazily from role: member configs. Re-spawning an explicit handle restores its history for the current lead session; dismissing an instance only removes it from the live roster.

---

Voice input

Click the mic button in the chat input to start voice input. Click again to stop. OpenAgentd uses the speech recognition built into your browser or app WebView and inserts the final transcript into the input for review — you still press Send manually.

There is no backend speech service, no faster-whisper dependency, and no speech.yaml setting. If the current browser/WebView does not expose speech recognition, the mic button is disabled with an explanatory tooltip. Platform privacy and cloud/local processing semantics are controlled by the browser or operating system speech service.

On macOS, voice input depends on the system speech service. If Siri & Dictation / Dictation is disabled in System Settings, Screen Time, or device management policy, speech recognition can fail with errors such as Siri and Dictation are disabled or Microphone permission check has failed. Enable System Settings → Keyboard → Dictation, then retry the mic flow.

---

Mobile app

OpenAgentd also has a Tauri mobile app for iPhone-class devices. It connects to a remote OpenAgentd server instead of bundling the Python backend locally, making it a companion client for the machine already running openagentd start --lan --key or another reachable server.

Mobile uses the shared backend-connection flow, the same React UI, and touch-first adaptations such as safe-area-aware layouts, keyboard-aware composer spacing, and single-pane navigation. See documents/docs/mobile.md.

---

Scheduler

Create tasks that run on a schedule or fire once at a specific time:

• Cron — standard five-field cron expressions
• Interval — every N seconds, minutes, or hours
• At — one-shot at an exact datetime

Tasks appear in the /scheduler panel. Pause, resume, or trigger them manually from the UI or via the REST API.

---

Observability

OpenAgentd exports OpenTelemetry spans to local JSONL partitions and serves a built-in dashboard at /telemetry:

• Summary — token usage, error rates, latency distribution, model breakdown
• Trace explorer — full span waterfall per session, filterable by date range
• Prometheus endpoint — /metrics for external scraping

No external collector required. All data stays on your machine.

---

Skills

Skills are .md files that inject domain-specific instructions into an agent's context on demand. They ship separately from agent configs, so one skill can be reused by any agent.

Builtin operational skills:

Skill	Purpose
self-healing	Agent edits its own config (model, tools, MCP, image/video settings)
mcp-installer	Install new MCP servers from the UI or by description
skill-installer	Install new skills from a URL or from scratch
plugin-installer	Install agent plugins

Add your own by dropping a SKILL.md file into {config_dir}/skills/{name}/ or via the /settings/skills UI.

---

MCP servers

OpenAgentd starts with an empty mcp.json. Add Context7 or any other MCP server via the /settings/mcp panel or by editing mcp.json directly. Changes are hot-reloaded without a restart.

{
  "servers": {
    "my-server": {
      "command": "npx",
      "args": ["my-mcp-package"],
      "env": { "API_KEY": "${MY_API_KEY}" }
    }
  }
}

---

Sandbox and permissions

Filesystem sandbox — A denylist blocks access to OpenAgentd's own data, state, and cache directories. Add your own glob patterns (**/.env, **/secrets/**) in sandbox.yaml. Changes take effect immediately, no restart needed.

Permission system — By default, tools auto-approve and log. Switch to interactive mode to block on sensitive operations and reply per-request with once, always, or reject. Permission decisions are persisted and replayed across turns.

---

Documentation

Full documentation index: documents/docs/index.md.

Getting started

Section	Contents
Features	Canonical, version-cited catalogue of every user-visible feature. Source of truth for slides, docs, and comparisons.
Roadmap / issues	Short planning note with links to GitHub issues for feature requests, future work, and known issues.
Install	Desktop app first (macOS/Linux); CLI/uv/pipx/pip, source.
Migration	Move setup from OpenClaw, Hermes, Claude Code, Codex CLI, or older OpenAgentd installs
CLI reference	Every openagentd subcommand
Slash commands	Reusable /name prompt templates; reuses your opencode command library
Shell commands	Opencode-style !command sends that run directly through the shell tool
Configuration overview	Hub — links into the focused subpages below
Environment variables	Settings fields, provider keys, optional extras
Paths & XDG roots	DATA / CONFIG / STATE / CACHE / WORKSPACE
LLM providers	Every registered prefix, OAuth flows, capability YAML
Agent files	.md frontmatter schema, validation, editing workflow
Built-in tools	Filesystem, shell, web, multimodal, scheduling, tasks, notes, team
LSP diagnostics	Coding-mode language-server feedback on edits, server selection, lsp: config
Skills	SKILL.md format, builtin skill catalog
Sandbox & permissions	Denylist paths, user sandbox.yaml, permission services
Comparison	How OpenAgentd compares to Claude Code, Codex CLI, Cursor/Windsurf, Aider, opencode
Troubleshooting	Common desktop-app and CLI/server issues

Architecture & internals

Section	Contents
Architecture	C4 diagrams, in-memory SSE streaming, SSE event protocol
Agent engine	Loop, hooks, tools, teams, plugins, context, summarization
Lazy team members	Spawn/dismiss member instances, blueprint#N handles, history restore
API reference	HTTP endpoints, SSE events, file handling

Operations

Section	Contents
Logging	App log + per-session JSONL, rotation, console format
Observability	OTel spans, DuckDB-backed /api/observability/*, /telemetry UI
Desktop distribution	Tauri v2 shell, Python sidecar, token auth, release pipeline
Mobile app	Tauri mobile shell for remote OpenAgentd API servers, backend connection flow, and touch-first layout details
Title generation	LLM-generated session titles, SSE event, config

Frontend (web/)

Section	Contents
App chrome	Shared header, platform detection, Tauri drag, macOS overlay
Coding sessions UI	Coding session restore, workspace sidebar pagination, reload/error handling
Workspace Files	Session files, coding Files & Diff, previews, downloads
Todos popover	Task board, assignments, claims, dependencies, live updates
Tool rendering	Tool call/result UI and custom renderers
Chat input & queue	Consecutive message queuing, PendingMessageQueue
Voice input	Client speech recognition and transcript insertion
Mobile layout	Phone-first responsive design — breakpoints, safe areas

Contributing

Section	Contents
Guidelines	Dev commands, code style, testing patterns, GitHub conventions
Team testing	Manual smoke-test recipes for the multi-agent team

---

Star History

---

Contributing

See CONTRIBUTING.md for setup, workflow, and PR guidelines.

Security

See SECURITY.md for the trust model and how to report vulnerabilities.

License

Apache License 2.0. Free for personal, research, and commercial use.