What can I help with?
Ask anything, run commands, explore files, or manage your scheduled tasks.
Approval required
Clarification needed
Pick a choice, or type your own answer below.
Terminal
Terminal
Drop files to attach
Select a skill
Pick a skill from the sidebar to view its contents, or create a new one.
Select a memory section
Pick a section from the sidebar to view or edit its contents.
Knowledge
Loading…
LLM Configuration
Knowledge Assistant
Configure the AI model that powers your global Knowledge store and RAG queries. Runs separately from your main conversation model.
Choose the AI provider for your Knowledge Assistant.
Fast, cost-effective models optimized for RAG and knowledge indexing.
Your API key is stored securely and never sent to Amy servers.
Select a scheduled job
Pick a job from the sidebar to view its details and runs, or create a new one.
Board
Select a task
Pick a task from the sidebar to view its details, runs, comments, and logs — or create a new one.
Select a space
Pick a space from the sidebar to view its files and settings, or add a new one.
Select an agent
Pick an agent from the sidebar to view its instructions, skills, connections, memory, and safety settings — or create a new one.
Current task list
No active task list in this session.
Usage Analytics
Loading...
Logs
Choose a log file to view recent lines.
No log lines yet.
Plugin
Conversation
Appearance
When enabled, the workspace / file browser panel opens automatically with each new session. You can still close it manually at any time.
When enabled, a Todos tab appears in the workspace panel (right side). The sidebar Todos panel remains available regardless.
Show floating Start and End buttons while reading long session histories.
When enabled, older messages load automatically as you scroll upward. When disabled, use the older-messages button.
Compact Worklog groups supporting activity into a quiet summary. Transparent Stream keeps each thinking/tool event as its own chronological expandable row.
When enabled, the view scrolls to the bottom as new tokens stream in. When disabled, you control the scroll position manually.
When enabled, bold, italic, links, and other markdown in your own messages are rendered. Off by default; fenced code blocks and math always render regardless of this setting.
When enabled, a profile switcher button appears in the top-left app titlebar so you can change profiles from any tab. Off by default; the composer footer always has a profile switcher regardless of this setting.
When enabled, new Worklog details start expanded so tool, thinking, and progress cards are visible without an extra click. When off, Worklog details stay folded by default; per-turn manual collapse/expand choices still win.
Choose which tabs appear in the sidebar and rail. Drag chips to reorder them. Chat and Settings are always visible.
Toggle which controls are shown in the chat footer. Reordering is not supported.
These controls do not always appear, depending on viewport, mode, or runtime state. You can still toggle them off if needed.
Preferences
Used for new conversations. Existing conversations keep their selected model.
Side-task routing for vision, compression, title generation, etc. "Auto" uses your main chat model.
Hide the three default suggestion buttons on the empty new-chat screen to avoid accidental taps.
Experimental. Renders only the visible portion of very long chat transcripts (over 80 messages) to speed up scrolling. Off by default because it can cause scroll flicker on long sessions with large tool-call blocks. Turn it on to try it; leave it off for the full transcript to render at once (and so the browser's Find (Ctrl+F) matches the whole conversation).
Flips alignment of chat messages and the composer input for languages like Arabic or Hebrew. Affects only the chat area — sidebar and other panels stay left-to-right.
Play a sound when the assistant finishes a response.
Show a speaker button on each assistant message to read it aloud using your browser's speech synthesis.
Automatically speak each new assistant response when it finishes. Pauses when you start typing.
Show the voice-mode button (audio waveform) next to the dictation mic. Lets you speak naturally — Amy reads replies aloud. Requires a browser that supports both speech recognition and TTS.
Record and send the original audio file to the agent instead of converting it to text first. The agent can then transcribe it or process the raw audio (emotion, background noise, custom STT). Like Telegram's voice message behavior.
Choose speech engine. Edge TTS uses Microsoft neural voices via the server.
Preferred voice. Populated from your browser's available voices.
1.0x
1.0
Show a system notification when a response completes while the app is in the background.
Displays input/output token count below each assistant reply. Also toggled with
/usage.Displays an ambient remaining-quota indicator (e.g. OpenRouter credit balance) in the composer footer. Default off. Only visible on wide displays (≥1400px) when enabled, to keep the composer uncluttered on laptop and standard desktop widths.
Show a desktop-only jump list for user questions in the current conversation. Off by default.
Displays tokens per second in assistant message headers while streaming and after a response completes. Off by default.
Fade newly streamed words in while the assistant is responding. Similar to OpenWebUI; off by default for maximum performance.
Expand the collapsed terminal panel automatically when a running command emits new output.
Self-hosted users can disable for transparency (not recommended for shared instances).
Controls how much metadata the session list shows in the left sidebar.
Maximum number of active conversations that can be pinned in the sidebar. Default is 3.
Automatically re-generates the session title based on the latest exchange, keeping it relevant as the conversation evolves. Requires an LLM title generation model to be configured.
Controls what happens when you send a message while the agent is running. Queue waits for the current task; Interrupt cancels and starts fresh; Steer injects a mid-turn correction without interrupting (falls back to interrupt when the agent is not yet cached or the stream has ended).
Compress conversation history when context usage exceeds this percentage (5–90). Default: 50. Lower = compress sooner; higher = allow more context before compressing.
Show conversations from CLI, Telegram, Discord, Slack, and other channels in the session list. Click to import and continue.
Supported channels: Telegram, Slack, Discord, WeChat, Zalo
Not yet supported: Signal, WhatsApp — no gateway adapter available
Not yet supported: Signal, WhatsApp — no gateway adapter available
Surface cron job output as conversations in the sidebar. Only active when non-WebUI sessions are enabled. Defaults off; high-frequency jobs can flood the sidebar.
Show older Discord, Telegram, Slack, and Weixin sessions that were replaced by reset or compression.
Mirrors WebUI token usage to state.db so
hermes /insights includes browser session data. Off by default.Show a banner when newer versions of the WebUI or Agent are available. Runs a background git fetch periodically.
Keep WebUI update checks on, but hide Agent update notices and skip Agent update fetches.
Changes the What's New action from opening the raw diff first to generating a short, human-readable summary. The regular diff comparison stays available after the summary.
Display name for the assistant throughout the UI. Defaults to AmyBot.
Providers
Integrations
Loading...
Policy
Allow, confirm, or deny rules and rate limits the agent honors for this connector.
More connectors
Additional connectors live here so the main page stays focused on the most common workflows.
Chat Channels
Loading...
Loading...
MCP Servers
Manage MCP servers configured in config.yaml.
Search known tools across active MCP servers.
Tool inventory only uses already-known active MCP runtime data; the WebUI does not start or probe servers.
Plugins
Extensions
Extensions run in the WebUI browser origin and can access the same authenticated APIs as this session. Only load trusted local extension directories; disabling an extension takes effect after reload for already-injected assets.
Loading extension diagnostics…
System
WebUI: —
Agent: not detected
Enter a new password to set or change it. Leave blank to keep current setting.
Gracefully stops the local WebUI server. Useful if you launched via
./ctl.sh start and want to stop without switching to a terminal. You'll need to relaunch the server (./ctl.sh start or the native app) before the WebUI is reachable again.Status of the Amy gateway (Telegram, Discord, Slack, etc.)
Loading…
Help