IRIS runs on Even Realities G2 smart glasses. Open it, tap Start, and it works — no keys, no setup, and your first 3 sessions are free with everything on. It listens with you, puts answers on your lens while the moment is still live, and remembers everything afterward. Below is everything it can do — all of it already working out of the box, none of it required. Customize only if you want to.
Not a demo reel — the ordinary moments where a heads-up display with a memory quietly earns its keep.
Ask IRIS how your day looks. It reads your calendar and to-dos together and tells you what actually needs you.
Someone states a “fact” that decides the deal. IRIS checks it while they're still talking.
Attach your prep notes before the meeting. When the question comes, the numbers are already on your lens.
Instructions, dosages, follow-ups — you just listen. The summary and the to-dos are waiting when you leave.
A negotiation is just a conversation where the other side knows more. IRIS levels it.
Your partner's family switches to their first language. Live translation streams onto your lens — you stay in the conversation.
Classes, sermons, seminars — IRIS keeps the key points and quiz-ready notes so you can be present in the room.
Every session becomes memory. Ask across all of them like you'd ask a friend with perfect recall.
IRIS knows who said what — and who they are the next time you meet them.
When your calendar has a meeting with someone IRIS knows, the Today screen surfaces what she remembers about them, recent history and open commitments, right under the event.
Turn Home into a lens hub and use IRIS hands-free: ask a question out loud and read the answer on the lens, glance at your agenda or to-dos, or command your connected agent — all from the ring. Pro, beta.
Built for deaf and hard-of-hearing wearers: the words being spoken around you, live on the lens with speaker labels — and nothing else. No cues, no AI, no recording, nothing saved.
IRIS listens with you and surfaces facts, answers, and advice the moment they help — visible only to you.
When someone asks something, IRIS hears the question mark and puts the answer on your lens before the pause gets awkward.
Enroll your voice once and transcripts, summaries, and action items name the speaker instead of guessing. Missed the prompt? Tag speakers on any past session later.
Mid-session, tell IRIS what matters right now — "only flag numbers", "we switched to Spanish", "stop cues for 10 minutes".
Quick actions are custom instructions on tap-able chips — "Summarize so far", "Translate that", "Draft a reply" — whatever you teach IRIS.
Too talkative? Too quiet? Adjust live from the recording screen — or pause cues entirely and use IRIS as a pure note-taker.
Your temple bar is the remote: tap, swipe, and double-tap cover dismissing cues, browsing history, paging notes, and switching modes.
Scroll up on the ring mid-conversation to open your quick actions right on the lens — pick one, and its answer appears where you're already looking.
Type or attach prep before a meeting — names, questions, files — then scroll up on the ring and open Prep notes to glance at them mid-conversation.
Your script on the lens, following your actual voice — not a timer. Someone interrupts with a question? It holds your place, marks where you were, and cues the answer. Afterwards the transcript is saved beside your script.
20 templates — 1:1s, lectures, sales calls, interviews — or describe your own format and IRIS builds the template for you.
Every session has a chat. Ask what was decided, or say "the title's wrong" — IRIS proposes the correction and you tap Apply.
Yesterday's meeting, on your glasses, while you walk to today's. Swipe the ring to page through.
Sessions propose memories worth keeping — people, preferences, facts. Saved memories sharpen every future cue and answer.
Promise to "ask Dana about the contract"? Next time you're in a session, IRIS quietly resurfaces it on your lens.
Tell IRIS what to remember — it sorts tasks into Home, Work, Study, Health and Personal, and spots to-dos in your conversations too.
Once a week IRIS reads back through your sessions: themes, open action items, and what slipped through.
Folders, merge, export, regenerate, transcript cleanup — the session list is a real archive, not a junk drawer.
Tap anyone in your Memory to see everything IRIS knows about them in one place: the facts you've saved, every conversation they were part of, and any open commitments that mention them.
After a meeting, IRIS writes a ready-to-send recap with the agreed next steps. Review it, tweak it, and send it from your own email or copy it anywhere. IRIS drafts it, you always send.
Sales calls, lectures, dates, negotiations — each mode has its own behavior, cue types, and chattiness. Swipe on the glasses to switch.
"Help me at a networking event — remember names, suggest icebreakers" → a full mode, generated, editable, yours.
When the conversation needs fresh facts — prices, news, flight status — IRIS searches the web mid-session and cites what it finds.
When IRIS answers a question using a live web search, she lists the sources right under the answer, so you can tap through and see for yourself where it came from.
Turn the heads-up display off to use IRIS as a phone recorder. It captures from your phone's mic with the lens dark, and your transcript, summary, and cues wait for you in the app. Good for meetings and note-taking when you'd rather the glasses stay dark.
Auto-save summaries straight into Obsidian — and retro-save any old session with one tap.
Browse Obsidian notes on the lens, and pin any note as context so IRIS cues from what you've written.
Replace the built-in Even AI with IRIS — same wake word, a much bigger brain, and it remembers. Ask anything, capture thoughts with the app closed, and recall from your IRIS memory and your recent recording summaries, all on the lens. Beta: capture and recall are dependable; live web answers (news, weather, scores) can still be unreliable.
Run Hermes Agent? Connect it and talk to your own agent from IRIS — its memory, its tools, its skills — replies streamed from your server.
Auto-export every processed session as a real Google Doc — headings, bullets, action-item checkboxes.
Every core feature works on Basic with your own API keys. Pro removes all setup — Claude and speech recognition are included — and unlocks the extras. 3-day free trial, cancel anytime.
| Feature | BasicBring your own keys · free | Pro$14.99/mo or $149.99/yr |
|---|---|---|
| Setup & AI | ||
| API keys to get started | You provide Claude & speech keys | None — fully managed |
| AI & speech-to-text usage cost | You pay the providers directly | Included |
| Live web search for cues | Your Tavily key or SearXNG | Included |
| In the moment | ||
| Live captions on the lens — accessibility, nothing saved | ✓ | ✓ |
| Live cues, answers & fact-checks on your lens | ✓ | ✓ |
| Live transcription & on-lens translation | ✓ | ✓ |
| Speaker ID — who said what | ✓ | ✓ |
| Modes, quick actions & ring controls | ✓ | ✓ |
| Prep file attachments per mode | 1 file | 5 files |
| Generate a mode from a description | — | ✓ |
| Presentation companion — notes & timer on lens | — | ✓ |
| Teleprompter that follows your voice & handles interruptions | — | ✓ |
| Home as a modular tool hub (beta) | — | ✓ |
| Narrow cue column — discreet, glance-like reading | — | ✓ |
| Hands-free lens hub — ask, agenda, to-dos & agent on the glasses (beta) | — | ✓ |
| Afterwards | ||
| Session summaries, memory & ask-a-session chat | ✓ | ✓ |
| Tag speakers on past sessions | ✓ | ✓ |
| Weekly review | ✓ | ✓ |
| Summary templates | 2 built-in | 20 + build your own |
| Session audio recording & playback | — | ✓ |
| To-do list, calendar, daily briefings & reminders | — | ✓ |
| Connected | ||
| Google Drive export | ✓ | ✓ |
| Google Tasks two-way sync | — | ✓ |
| "Hey Even, remember…" quick capture, app closed (beta) | — | ✓ |
| Obsidian sync & browse-on-glasses | — | ✓ |
| Hermes agent bridge | — | ✓ |
| Cloud settings & data backup | — | ✓ |