OcuClaw

Timeline

OcuClaw, the story so far.

The releases, fixes and milestones that shaped OcuClaw.

Stable release Hardening / fixes News

Release timeline — newest first

30 Sep 2026 News

Featured by Even Realities.

Even Realities’ new Developer series opens with OcuClaw. Watch the episode.

29 Sep 2026 v2.1.1 · Stable release

OcuClaw 2 is here.

Our biggest update so far. OcuClaw V2 brings Hermes Agent support, interactive displays, silent typing and a new home for your agents. Meet Speck, follow work from your glasses, and make OcuClaw your own.

  • Hermes, alongside OpenClaw — chats, approvals, models and skills now work with Hermes Agent too. Its kanban board lets you follow several agents at once, review their work and answer when they need you.
  • LiveUI — your agent builds checklists, choices, readers and graphics on your glasses. Save useful interfaces as Tasks and run them again.
  • Silent input — compose messages with swipes and taps on your ring or glasses. We built a custom word model trained on real conversations and optimised to minimise the inputs needed per word. It learns your vocabulary and uses the message you’re answering for context, while Jev by TypeSafe AI helps rank likely next words. Your agent can also suggest whole replies.
  • A new Home — see your agents, spot who needs you and switch between chats. Add agents for different jobs, with Light, Dark and System themes.
  • Your controls, your way — pin skills, Tasks, agents and Saved Prompts to the contextual menu or a temple gesture.
  • Meet Speck — the Alive engine brings Speck to life, reacting as your agent thinks, uses tools and works through the Board.
  • Smoother reading and voice — replies reach the glasses sooner, Natural Text Flow follows your reading pace, and interrupted voice input stays available for review.
  • Easier setup — guided installation and pairing cover local and Cloudways hosts, with clearer progress, scannable QR codes and better recovery when a step needs attention.
18 Sep 2026 News

OcuClaw now runs on System One.

TypeSafe AI granted OcuClaw early access to System One Models, a new class of model built to make fast, structured decisions. Jev, the first of them, is available today in Silent input, where it ranks likely next words using the context of your message. Coming soon to LiveUI: Jev joins the pipeline that builds the interfaces on your glasses, making screens even faster to appear and lighter on the session, so the agent holds on to more of your conversation.

08 Sep 2026 News

Partnered with Cloudways.

OcuClaw has partnered with Cloudways to offer the best cloud OcuClaw experience. Run OpenClaw or Hermes on a managed server so your agent stays online when your computer sleeps, with guided setup from install to your first message on the glasses. Need help? Cloudways staff answer in their own channel in the OcuClaw Discord.

01 Sep 2026 News

Security at the gate.

Vercel deepsec now checks every OcuClaw deployment, combining frontier models to continuously scan the client, plugin and infrastructure. Any flagged vulnerability blocks the release until cleared. OcuClaw 2.x will be the first release cleared by deepsec.

21 Aug 2026 V1.4.0

Tap, hold and talk.

OcuClaw 1.4.0 lands on SDK 2.9.0 with a customisable contextual menu, long-press push-to-talk and an optional live voice waveform.

  • Customizable menu — use the new tap-then-long-press contextual menu, then customise and rearrange its actions to suit how you use OcuClaw
  • Push-to-talk voice — long-press to start and finish transcription without opening the menu
  • Waveform visualizer — enable an optional waveform that reacts to your voice in real time while you transcribe
04 Aug 2026 News

Featured by Fly.io.

The company behind Sprites, Fly.io, spoke with us about our Fleet Control Center: the agentic development factory behind OcuClaw. Read the interview.

31 Jul 2026 News

Finally, a website.

09 Jul 2026 v1.3.7

Boot watchdog and clearer errors.

A hardening release: nothing hangs silently anymore — stuck loading screens get a watchdog, voice failures get real error labels.

  • Boot-splash watchdog — a stuck loading screen steps aside instead of eating taps
  • Voice-error header labels — failures name themselves, including "No Soniox credits"
  • In-app bug reports — fixed a timezone bug in the reporter itself, so anyone west of UTC can file again
05 Jul 2026 v1.3.6

The big stable.

The first public stable since v1.2.4. Setup is agent-led now, the phone and glasses apps were rebuilt, voice is faster, and recovery actually tells you what went wrong.

Setup, phone & recovery

  • ClawHub and OcuClaw-Assist give your agent one guided lane to install, upgrade and repair OcuClaw.
  • The rebuilt phone app adds per-setting help, a full Sessions browser and one-tap diagnostic reports.
  • Smarter reconnect, typed connection errors and a hardened relay turn silent dead ends into recoverable states.

Glasses, voice & Neural

  • A new glasses renderer fills pages cleanly, streams with less flicker, and text flows continuously instead of paging.
  • Voice gains Cartesia Ink-2, Soniox v5 and a faster, more reliable capture-to-response handoff.
  • Topic Distiller, Emoji Reactor and Pace Modulator let the agent shape session titles, expression and reading rhythm.
05 Jul 2026 News

OcuClaw is live on ClawHub.

Stable installs are now one command — openclaw plugins install clawhub:ocuclaw. Updates ride the same rail.

03 Apr 2026 v1.2.4 · Launch

Where OcuClaw started — your agent in your field of view.

Where it all started. OcuClaw brought OpenClaw to Even G2 — voice, streaming replies, live tool activity, approvals and persistent sessions, all in your field of view.

A conversation, not a notification

  • Voice in. Streaming answers out. Long replies flow through glance-sized pages paced to your reading speed.
  • Twenty-two screens cover conversations, sessions, images, models, skills and settings — a real OpenClaw client, not a remote prompt box.

The agent stays visible

  • Thinking summaries and live tool activity show what is happening while the agent works.
  • Approval requests land on the glasses with Allow Once, Allow Always and Deny — action stays in your control.

Join in

Follow the beta channel.

Every update above lands in the Discord first. Come try the builds before they're stable.

Join the Discord Set up OcuClaw

Release archive

OcuClaw 2.1.1

OcuClaw 2 is here. Our biggest update so far: Hermes Agent support, LiveUI, Silent input, a new Home and Speck.

🚀 New: Setup and pairing

  • Agent-guided setup — Ask your agent to set up OcuClaw and it walks you through installation, pairing and your first message. If you stop partway, it picks up where you left off.
  • Fewer approvals — Setup gets what it needs from the plugin directly, so you approve fewer commands.
  • End-to-end check — Setup sends a test message, checks reply delivery to your glasses, and finishes with a welcome card you double-tap. If your model fails, it tells you why and the phone’s setup bar turns red with a Help button.
  • QR pairing — openclaw ocuclaw pair shows a QR code. Scan it and check that the same four words appear on both sides. The code expires after two minutes and contains no secret.
  • No relay token to copy — The plugin creates its own credential. Existing tokens keep working.
  • Voice and Even AI — Offered on Home once setup is done. API keys go into private popups and are saved on your host, never in chat.

📌 Old setup skill — If you installed the separate ocuclaw-assist skill, remove it with openclaw skills remove ocuclaw-assist. The setup assistant now comes inside the plugin, and the old copy would keep you on an outdated guide.

☁️ New: Cloudways

Run your agent on a Cloudways server so it stays online.

  • One command — openclaw ocuclaw cloudways setup or hermes ocuclaw cloudways setup walks through numbered steps from installation to your first message on the glasses. No sudo needed.
  • Checks along the way — Confirms OcuClaw loaded, helps you sign in to a model, shows the Tailscale approval link as a QR code, and survives a container restart. ChatGPT subscriptions work too.
  • HTTPS certificates — Waits up to five minutes for the certificate, with a progress update every 30 seconds.
  • Connection help — If the phone can’t reach your agent, a card lists likely causes with Help and Try again.

🪽 New: Hermes support

OcuClaw now works natively with Hermes Agent as well as OpenClaw.

  • Chats, approvals, models and skills all work on Hermes. Each profile has its own chat.
  • Multiple agents — Choose which Hermes profiles OcuClaw uses, create new ones from the phone, and switch between them from the glasses.
  • Settings › Hermes — Manage pending approvals, approval rules, scheduled jobs and skills, and check health and usage.
  • Continue desktop chats — Pick up a chat from Hermes Desktop, the CLI or the TUI on your glasses. One device drives at a time, with Take over to switch.
  • Hermes Desktop — Optional OcuClaw theme and G2 Pulse in the title bar, showing connection, battery and recent activity, locally or over SSH.
  • Phone tools — With access enabled, your agent can read your location and glasses battery, and name the chat.
  • Setup helper — Installs, pairs and updates in place with one gateway restart. Works on Windows too.
  • Plan progress — When Hermes works through a plan, the glasses header shows the current step, such as “Step 2/3 · Run the tests”. Change this under Settings › Display › Agent activity.
  • Hermes 0.21.5 certified.

🗂️ New: Hermes kanban

Give several Hermes agents their own jobs and track them all on one board.

  • Lanes and cards — Needs you, Running, Blocked and more. Open it with the Board button in the top bar.
  • From the phone — Create cards, comment, Approve or Request changes, Retry, reassign, and open a card’s files.
  • From the glasses — Add Board to your menu to browse lanes, Approve on the wheel, or speak a change request. If Hermes’ kanban tools are enabled for OcuClaw, you can create cards by voice.
  • Notifications — Your glasses and phone tell you when a card you’re watching needs a review or an answer, or fails. Quiet hours hold notifications until they end.
  • Conflict checks — Every action checks that the card hasn’t changed since you saw it.
  • Board animations — Speck acts out kanban work such as filing, checking and unblocking cards, and Board lanes animate as you swipe.
  • Compatibility — Tested on Hermes 0.21.1 to 0.21.5. Newer 0.21 builds get Board if a compatibility check passes.

🧩 New: LiveUI

Your agent builds interactive interfaces on your glasses. Move through them with swipes and taps on the ring or glasses temples.

  • Readers, lists, checklists and pictures, plus graphics such as metric tiles and sparklines.
  • Tasks and Templates — Save a request to run again later, such as a morning briefing, or save a layout to reuse. New Tasks need your OK on the phone.
  • Library — Find your saved items on the phone and glasses, and pin them to the glasses menu.
  • Permissions — Websites need your OK once. Turning LiveUI off keeps your Library.
  • Short connection drops — A LiveUI page survives a phone connection drop of up to 12 seconds. A × appears while the link is down so you know the page may be stale.

👓 New: Speck

Meet Speck — A small drawn character in the top left of the display, brought to life by our new Alive engine.

He follows what your agent is doing: listening while you talk, thinking, reading and editing files, searching the web, handing work to sub-agents, waiting on your approval, and building your LiveUI interfaces.

Prefer emoji? Switch to Emoji in Display settings.

⌨️ New: Silent input

  • Type without talking — Swipe and tap on your R1 to spell words with three-way predictive letters, or pick exact letters.
  • Our own word model — We trained a custom word model on real conversations with one goal: the fewest inputs per word. It learns the words and word pairs you use, brings back words you’ve spelled before, handles contractions such as “don’t”, and uses the message you’re replying to for context.
  • Jev by TypeSafe AI — Helps rank likely words so the one you want appears sooner. Add a TypeSafe key on the phone to enable it.
  • Suggested replies — Your agent can suggest whole replies. They open for review and never send on their own.
  • Expanded vocabulary — Spell exactly recognises about 150,000 more words, including names, places and technical terms.
  • Punctuation — A comma takes two moves, and Review can add the “?” or “.” for you.
  • My words — Keep favourite words in and hide ones you never use.
  • Temple tap warning — Five quick taps on one temple restart the G2, so Silent input warns you when a fourth quick tap lands on the same temple. Every tap still counts.

✋ Menus and temple controls

  • Redesigned tap-then-hold menu — New session, Session, Prompts, Model, Change agent, Skills, Silent input, LiveUI, Settings and Compact context.
  • Customise your menu — Up to ten rows from built-ins, agents, skills, Tasks, Templates and Saved Prompts, with a live preview on the phone.
  • Temple actions — Double-tap either temple to open Sessions, LiveUI or Agents, or run a skill, Task, agent or Saved Prompt.
  • Wide menus — The Agent, Settings, Read Speed, Thinking, Skills, Model and Session pickers show the selected item and its value side by side.
  • Saved Prompts — Write up to 30 on the phone and send them from the glasses.

📖 Replies and reading

  • Natural Text Flow — Replies appear at your reading speed, with brief pauses at commas and sentence endings.
  • Words sooner — Replies show up faster, and long replies and long chats stay responsive as they grow.
  • Two-part replies — When your agent writes, runs tools, then writes more, the first part stays on the display.
  • Scroll active replies while you read. Double-tap jumps back to the head.
  • Activity display — Choose how much of your agent’s work you see: Off, Status bar, Show the work or Narration.
  • Questions on the glasses — Answer with one choice, several choices, or by voice.

🎙️ Voice

  • Tap or Hold — Tap mode can be Manual, where you review before sending, or Automatic, which sends when you stop. Hold replaces the old Listen mode.
  • Holds up to three minutes — Words already transcribed are kept, even when interrupted.
  • Live transcription — Words appear as you speak and get corrected in place.
  • Custom vocabulary for Cartesia, and speaker labels on Soniox.
  • Faster start — Push-to-talk starts more quickly.
  • Your words are kept — If voice fails mid-take, or you let go before it connects, what you said waits in Review instead of being lost.
  • Fewer hangs — A stuck mic recovers on its own, push-to-talk after “Hey Even” is no longer muted, and a late result can’t send your message twice.

📌 Voice defaults to Hold — If you never switched to Tap, voice now uses Hold: hold to talk, let go to stop. You can switch to Tap in Voice settings.

📱 Phone app

  • New Home — Built around your agents. Avatars show who’s working, has unread messages or needs you. Chats appear as chips, with model, thinking and context below.
  • Create and edit agents from the phone.
  • Sessions — Tap a chat to pin, rename, copy, hide or delete it. Search transcripts, and type / for commands.
  • Dark mode — Choose Light, Dark or System under Settings › Display › Appearance.
  • OcuClaw font throughout the phone app.
  • Info popup — What’s new, Report a bug and help.

📶 Connection and network

  • Less data — On Android, replies stream compressed, using roughly one-fifteenth of the data for a long reply in our tests. That helps replies keep up on a weak signal.
  • Steadier link — The phone keeps a healthy connection open more often and detects a dead one in about 10 seconds, down from about 30.
  • Reconnects keep your place, and the glasses display stays off while it sleeps.

🛡️ Reliability and polish

  • iPhone — OcuClaw starts and keeps running when the Even app isn’t in the foreground.
  • Android — Opening OcuClaw from the glasses while the Even app is in the background now starts up properly.
  • Smaller fixes throughout — Display, voice, approvals, reconnects and startup all receive fixes.

Release archive

OcuClaw 1.3.7

Fix-only release (app + plugin).

🩹 Fixed

  • Loading screen could hang forever on startup and block all taps.
  • "No Soniox credits" warning wasn't shown when your Soniox account ran dry — listening just ended silently. Proper warnings now surfaced in the status bar for credits, key and rate-limit errors.
  • In-app bug reports never worked west of UTC-0 (all of the Americas) — fixed via the plugin, so it starts working as soon as your plugin updates, no app update needed.

Release archive

OcuClaw 1.3.6

Everything that landed across the beta cycle, rolled up into the first stable, publicly-installable release since 1.2.4.

🧭 New — guided setup skill. Run openclaw skills install ocuclaw-assist, then ask your OpenClaw agent to install, upgrade or repair OcuClaw — it walks you through each step.

🦞 OcuClaw is now on ClawHub — openclaw plugins install ocuclaw just works.

Grab a coffee, this is a big one. ☕

🧠 New: Neural Modules

The agent can now reach past a plain text reply and shape how the response feels on your glasses.

Neural Emoji Reactor — lets the agent nominate a status-bar emoji for specific spans of its streaming response, swapping it in sync with what you're reading. Reserved for genuinely impactful moments; the usual status-bar behaviour runs the rest of the time. Toggleable in the Display tab.

Neural Pace Modulator — lets the agent dial the reveal speed of specific phrases as it streams: <dwell> to slow a line down so it lands, or <skim> to rush past a recap that doesn't need to breathe. Everything else reveals at your normal reading pace and it combines cleanly with the Emoji Reactor.

Neural Topic Distiller — the agent gives each session a short title from context and quietly renames it as the discussion shifts. Ask it to rename anytime, or edit titles yourself; a manual rename always wins and locks the auto-titler out of that session. It runs as a small background agent after your turn, so it never bloats your prompt or derails a reply.

🛠️ New: Agent Tools

Hardware Info — a new agent tool that surfaces live G2 hardware state, starting with battery percentage, so the agent can factor your glasses' status into its replies.

🎙️ Voice & Speech-to-Text

Cartesia Ink-2 is now a second STT provider — pick your speech-to-text engine in Voice settings. Ink-2 is purpose-built for real-time voice agents, with ~100 ms to a final transcript, built-in turn detection and strong capture of structured data.

Soniox default model → stt-rt-v5 — the retired v4 is auto-migrated. Accuracy improves across 60+ languages, noisy audio, accents, interruptions and overlapping or mixed-language speech.

  • Endpoint Sensitivity tuning on glasses and WebUI controls how eagerly the mic decides you've stopped talking.
  • Conversation Context primes the recognizer with your agent name and session title.
  • The entire voice pipeline is faster and more reliable from capture through stream-start.
  • Voice-latched lifecycle keeps the spoken user message visible through capture, commit, thinking and stream start.
  • Configurable voice auto-timeout, with a glasses selector and WebUI controls.
  • Continue-listening setting for chained Even AI voice turns.

🗂️ Agents & Sessions

  • Per-mode and per-session Agent selection — choose which OpenClaw agent runs for each mode and each session.
  • WebUI per-session agent switcher — swap to a different specialist mid-work.
  • Session multi-select agent filter — filter your session list by one or more agents.
  • Faster Sessions tab — history is cached on-device, reconnect syncs by diff and transcript search can auto-run.
  • Better mobile session search — the fixed search row uses a native mobile keyboard and keeps focus after the first character.
  • Active turn controls — stop, replace or steer an in-flight agent turn.
  • New WebUI Sessions browser — OcuClaw and Even AI tabs, Preview, Pin, Rename, Delete, bulk actions, live search and pinned favourites.

👋 First-launch Experience

  • A new welcome screen stays on the glasses until you connect, with an animated OCUCLAW wordmark, scanline loader and tagline type-in.

💬 Messaging Screen Redesign

  • New text rendering engine — the messaging screen reliably fills more of each page before flowing to the next.
  • Faster, more stable, lower-flicker streaming, page-0 transitions, session switches and voice hand-off.
  • Voice and typed messages display with parity: the same first-chunk behaviour and reveal pacing.
  • Listen-resume holds until you've finished reading the streaming response.
  • Unified header and body rendering with independent truth models.
  • Assistant typing placeholder before the first response content arrives.
  • Manual-send overlay on the glasses after sending from the phone UI, remote control or voice.

🌊 Streaming Modes & Flow

  • Continuous mode scrolls the response line-by-line instead of jumping to a fresh page.
  • Dynamic flow eases the reveal in and subtly changes pace at paragraph endings.
  • Word-atomic reveal prevents mid-word line-wrap flicker.

🪟 Page Indicator & Scroll Bar

  • Scroll-bar marks response-start pages with · and continuation pages with ..
  • Fast-scroll v2 ramps page skips as you go; a one-second pause resets the streak.
  • The page indicator is pixel-stabilized and right-anchored.

😀 Agent Emoji

  • Speaking animation while text streams.
  • Idle heartbeats every 30–60 seconds and short mini-sequences every 2–4 minutes.
  • Show Agent Emoji master toggle, plus a separate Idle Animations toggle.

🎛️ Agent Modes

  • Fast mode toggle — opt supported models into faster output, per session.
  • Exec escalation mode toggle — choose whether the agent can request elevated exec access, per session.

🐞 Bug Reports & Diagnostics

  • One-tap Send Bug Report — add a note, review what's included, then send the bundle with a copyable reference.
  • Works while disconnected — builds a client-only diagnostics bundle with an offline export escape hatch.
  • Selectable history window — defaults to the last five minutes, with a Max option.
  • Safer contents — secrets are stripped before upload and Review shows exactly what's included.
  • Record a window — start and stop a short diagnostic capture from the WebUI with a live timer.

🧭 Glasses Skills Menu

  • Redesigned skills browser with a fixed selection box, descriptions, safe truncation and clean swipe or tap navigation.
  • Cleaner agent menu labels, with long names truncated safely and the default agent marked directly on glasses.

⚠️ Transient Failure Feedback

Short error labels now appear in the glasses header for about 3.5 seconds, then clear on your next listen, send, chunk or disconnect:

  • Assistant run: Out of credits, API key invalid, Rate limited, Provider unavailable, Provider timeout, Request failed, Run failed.
  • Voice: No Soniox credits, Soniox key invalid, Soniox rate limited, Soniox unavailable, Soniox key missing, Mic access failed, Voice disconnected.
  • Send: Image send failed, Image too large, Connection lost, Send failed.

The relay setup section auto-expands with a typed error for every connection failure — Invalid token, Bad relay address, Wrong server type or Cannot reach relay — and the Connect button changes to Retry.

🔐 Exec Approval

  • Approval screens honour server-provided allowed decisions, with improved reliability.

👓 Glasses Display

  • Root-menu double-tap opens the SDK exit prompt; the old Exit row becomes a Conversation shortcut.
  • Glasses rate-limit indicator — the header shows reset time when limits are hit.

📱 Phone UI, Theme & Settings

  • Brand-new phone app theme and layout, with icons across every setting.
  • Help button, Help modal and per-setting help across Display, Session, OcuClaw, Even AI and Voice.
  • Agent avatar or emoji across the phone UI, cached client-side.
  • Session token usage with a compaction button.
  • Even AI controls visibly disabled when Even AI isn't enabled — no more dead controls.
  • Improved iOS and Android keyboard handling.
  • Refreshed Info popup; Report a Bug is easier to find and Discord points at discord.ocuclaw.com.

📊 Provider Usage & Rate Limits

  • Adaptive hero card and secondary list in the OpenClaw Status pane.
  • Multi-auth provider-pool display suppresses the warning while a fallback is ready or rotating.
  • Rate-limit details and reset windows render in your local timezone.

🛡️ Reliability, Performance & Polish

  • Removed an O(history) layout cost so long replies stream without build-up stutter.
  • Thinking glyph paints about 40% faster on G2; glasses bridge writes are bounded so a stuck call can't hang rendering.
  • Page-0 paragraph-break freeze fixed, plus dozens of smaller streaming, voice, session-menu and display fixes.

🔗 Connection & Relay

  • Smarter reconnect remembers which addresses have proven good and re-seeds them on cold-start auto-connect.
  • Hardened relay worker adds backoff, watchdog, socket liveness, ping reaping, slow-client shedding, queue caps and connect timeouts.
  • Plugin and app versions are exchanged in the handshake; the WebUI detects a mismatch and pauses auto-reconnect.
  • Reconnect tolerates slower networks and high-latency Tailscale or cellular handshakes.
  • Honest “never connected” copy and restored per-error messages in the connection panel.
  • The 1008 reconnect storm and several first-connection or iOS first-start lockups are fixed.

📌 Migration note — Connection

The default Tailscale setup now uses dual-port: the relay moves to direct TCP wss://…:8444, while https://…:8443 stays for the Even AI endpoint. Existing users are walked through the swap by the installation support agent.

🚚 Distribution, Install & Onboarding

  • OcuClaw is now published on ClawHub — a bare openclaw plugins install ocuclaw resolves to ClawHub by default; npm remains the beta channel and fallback.
  • Minimum host version raised to OpenClaw 2026.6.9.
  • Guided setup lane — install openclaw skills install ocuclaw-assist and let your agent walk through install, upgrade and repair.
  • Config schema enforces relayToken and evenAiToken requirements at config-set time, with friendlier startup errors.

📌 Migration note — Even AI

If your config has evenAiEnabled: true without evenAiToken, updating will fail validation. Set the token, or disable Even AI before updating.

📌 Migration note — Tools profile

If openclaw config get tools shows a profile, grant plugin tools back with openclaw config set tools.alsoAllow '["ocuclaw"]' --strict-json, then restart the gateway.

⚙️ Behind the Scenes

  • Website launching soon. The release timeline, setup and support material were on the way.
  • Fleet Control Center — a web cockpit for remote development boxes with live glasses streams, input routing, live diagnostics, trace tabs, screenshot galleries and milestone notifications.
  • Debug-triage service — uploaded bug reports feed a secured dashboard and agent API for investigating bundles safely.
  • remote-debug skill — stand up a public, token-gated instrumented build for an external tester and watch the live boot or trace beacon.

Thanks to everyone who ran the betas and sent reports.

Release archive

OcuClaw 1.2.4

Where it all started. OcuClaw brought OpenClaw to Even G2 — voice, streaming replies, live tool activity, approvals and persistent sessions, all in your field of view.

💬 Messaging & Conversation

  • Bidirectional chat with OpenClaw agents on Even G2 glasses.
  • Real-time streaming response display with auto and manual scrolling.
  • Turn-based message pagination with a virtual pager.
  • Image attachments via camera capture or file picker.
  • Session history browsing and switching.
  • Slash command forwarding.

🎙️ Voice

  • Soniox speech-to-text with a warm lease system.
  • Two access modes: tap for direct glasses input and menu for ring-controlled navigation.
  • Automatic endpoint detection at speech end or manual commit mode.
  • Configurable endpoint-detection delay based on each STT model's capabilities.
  • Real-time transcription overlay on glasses.
  • Pressure-aware cadence throttling adapts overlay updates to streaming load.
  • Audio capture automatically pauses during agent thinking.
  • Pre-connect audio buffering queues up to 150 chunks before the STT session is ready.
  • Context-aware voice recovery handles mid-utterance, post-commit and transport interruptions.
  • Language hint configuration using ISO codes improves STT accuracy.
  • Listen-on-startup option.
  • Multiple STT models with per-model capability tracking.
  • Instant transcription erase and continue from a ring swipe, without breaking speech.
  • A separate Even AI voice mode detects the open OcuClaw app and can route wake-word utterances into the active session.

🧠 Even AI

  • Dedicated Even AI session routing.
  • Three routing modes: active OcuClaw session, persistent Even AI session and new Even AI session.
  • Custom system prompts, model selection and model configuration for Even AI requests.
  • Separate session and model defaults.

⚙️ Models & Skills

  • Dynamic model catalog from OpenClaw.
  • Per-session model and thinking-level configuration.
  • Reasoning and verbosity controls.
  • Skills browser with tap-to-invoke directly from glasses.

✨ Activity Status

  • Live header status with 15 emoticon indicators across 24 state and intent mappings, including thinking, tool execution, reading, browsing, fetching, subtasks, coordinating, sending, queued, streaming and global connection states.
  • Semantic intent classification for file reads, writes, edits, searches, Git, browser, network, agent, session and canvas operations.
  • Thinking summaries extracted from agent reasoning and displayed inline.
  • Smart tool-to-summary handoff preserves emoji context across transitions.
  • Priority queue with dwell timing creates smooth, flicker-free status changes.
  • Stale and duplicate activity guards use per-run sequence tracking.
  • Configurable activity visibility, tool-only or tool-plus-thinking mode, and normal or compact verbosity.

🔐 Exec Approval

  • Approval requests display full command metadata.
  • TTL countdown timer.
  • Decision queue for multiple pending approvals.
  • Allow Once, Allow Always and Deny actions.

📐 Text Rendering & Pagination

  • Custom glyph-metrics engine with awareness of narrow and wide character widths.
  • Emoji allowlist filters assistant text to more than 50 display-safe emojis for OcuClaw and Even AI.
  • Multi-pass text-density rebalancing removes sparse pages using sentence and paragraph boundaries.
  • Seam-snap post-processing refines page boundaries with orphan detection and cost scoring.
  • An automatically prepended readability prompt optimizes agent output for the small-screen display.

🕶️ Glasses Display

  • Twenty-two dedicated glasses screens across messaging, menus and settings.
  • Configurable display timeout for inactivity sleep.
  • Reading-speed controls use WPM to auto-advance streaming pages.
  • Stream-page delay and fast-scroll options.

📱 Phone UI

  • Responsive companion WebUI with tabbed settings.
  • Relay server configuration with a setup help card.
  • Near-full settings parity across Display, Session, OcuClaw, Even AI and Voice.
  • Debug section for advanced diagnostics.

🔗 Connection & Relay

  • WebSocket relay through the OpenClaw Native Gateway plugin.
  • Token-based authentication.
  • Automatic reconnection with a revision-tracked resume handshake.
  • Session retention with pruning and first-message caching.
  • Settings persistence and synchronization between phone and glasses.