Convene the room.
Stress-test the decision.
Cast a CFO, a port superintendent, a head of merchandising — anyone you need. Add real teammates. Watch the meeting unfold turn by turn, with an orchestrator deciding who speaks next.
Free 500 tokens on signup · No credit card required · "Build a meeting" auto-casts Q3 budget pressure-test
Q3 budget came in 8% over. Should we cut headcount or trim the new market launch?
Headcount is a one-time hit but recurring savings. Show me 3-quarter cash impact for both before we touch people.
If we cut, we lose two PMs we just hired. Cost us 6 months of momentum. Let's pressure-test the launch first.
Launch can be staged — soft-open one region, defer paid spend. Saves ~40% of the launch budget without killing it.
Where the synthetic boardroom meets the real one.
- Hospital throughput committees
- Container terminal ops
- Pre-market trading desks
- Sprint delivery huddles
- M&A diligence rooms
- Board strategy off-sites
Not a transcript. A decision.
Every huddle ends with a synthesised brief — decision, guardrails, targets, named owner. One artifact your team can act on.
Hold the date. Cut bulk export & saved views. Ship via flagged rollout, gated on the CVE patch.
CVE patch merged · 24h QA soak · auto-revert on >2% error budget burn.
Zero high-severity CVEs at GA · ≤1% rollback rate · regression green on the last two builds.
Eng Manager owns the release. SRE owns the rollback drill.
Four steps. Sixty seconds to set up.
- 01
Pick a kit
31 industries pre-built, or start blank.
- 02
Builder pre-fills
Role, tone, KPIs, do/don't rules.
- 03
Drop your docs
Personas cite your data, not the open web.
- 04
Convene
5 seats, one decision. ~6 minutes.
Try the persona power-search
The same fuzzy search, multi-skill filters, and guided builder you get inside the app — running live, right here.
Payer mix, DRG economics, and capital planning for a 600-bed system.
Restricted vs unrestricted, indirect rate, and audit compliance.
Compose tone, goals, and do/don't rules — exactly like the wizard inside the app.
One click, a fully-cast meeting
Each kit creates a meeting with the brief, context, and AI cast pre-loaded. Sign in and you're in the room.
Q3 budget pressure-test
Stress-test an over-budget quarter with finance leadership.
Pre-market trading floor huddle
15 minutes to the open — hold, hedge, or de-risk after an overnight macro shock.
AI feature launch readiness
Ship-or-slip review for a new LLM-backed product feature.
Holiday assortment review
Lock the Q4 hero items, price ladder, and promo cadence.
Asia–Europe blank sailings
React to soft demand on the headhaul without breaking customer commitments.
Sprint risk & delivery huddle
Mid-sprint hold-or-slip call with engineering, product, QA, and SRE.
Everything a real working meeting needs
Beyond a chat with one model. Convene gives every seat a persona, a memory, and the tools to actually do work.
Up to 8 seats
Mix yourself, real teammates, and AI personas in any combination.
230+ industry personas
Curated experts across 31 industries — finance, healthcare, IT, real estate, aerospace, government and more — fully editable.
Smart orchestrator
Auto-paces the meeting and picks who should speak next based on the topic.
Invite real humans
Share a link or email invite. Online / Away / Offline presence shows who's in the room.
Document context
Drop in PDFs, slides, images. Attendees read and reason over them automatically.
Tools & deliverables
Connect APIs and approve writes. Export decisions, summaries and artifacts in one click.
Pick your model
Gemini Flash for fast turns, GPT-5 or Gemini Pro for deep reasoning — per participant.
Built-in security
Login lockout, owner-scoped data, and host-only controls for guest seats.
Guided persona builder
Compose tone, goals, and do/don't rules with a wizard — save personas to your library and reuse them anywhere.
Power-search the library
Filter roles by industry, seniority, and multi-select skills. Fuzzy search with highlighted matches across names and instructions.
Bring the right expert to the table
31 industries, 230+ curated AI personas. Pick one, tweak the brief, drop them in.
Need something else? Every template is fully editable, and you can save your own.
From blank page to real conversation in 3 steps
- 01
Set the scene
Name your meeting, drop in the topic, and attach any documents the room should know about.
- 02
Cast the room
Pick AI experts from the template library, customize their persona, and invite real teammates by link or email.
- 03
Run the meeting
Speak as yourself, let auto-pacing orchestrate turns, then export decisions and deliverables when you're done.
Pay for what you actually use
One token economy across meetings, AI replies, and deliverables. Start free with 500 tokens.
Starter
- AI reply20–40 tokens
- Add AI participant20 tokens
- Create meeting30 tokens
- Generate deliverable100 tokens
Growth
- AI reply20–40 tokens
- Add AI participant20 tokens
- Create meeting30 tokens
- Generate deliverable100 tokens
Enterprise
- AI reply20–40 tokens
- Add AI participant20 tokens
- Create meeting30 tokens
- Generate deliverable100 tokens
Hosts pay for everything inside their meetings — invited guests never spend tokens.
Common questions
Getting started
What is CollabAI in one sentence?
It's a meeting room where you cast multiple AI personas (and optionally real teammates), debate a decision in chat, and walk out with a synthesized deliverable.
Do I need to install anything?
No. CollabAI runs entirely in the browser. Sign up, open a room, and you're live in under a minute.
Can I try it before paying?
Yes — every new account gets 500 free tokens, enough for a full first meeting with multiple AI seats and a deliverable.
Meetings & collaboration
Can I bring real teammates into a meeting?
Yes. Invite by email or share a link. Guests join with chat-only access and the host pays all tokens. Presence is shown as Online, Away, or Offline.
How many AI personas can join one room?
Up to 8 AI seats per meeting, each with its own persona, model, and system prompt. You can save kits to reuse the same lineup later.
Can the AI take real actions?
Yes — connect APIs and the room can call them. Writes can require host approval before they execute, so nothing ships without your sign-off.
Models, tokens & pricing
Which AI models can I use?
Each AI participant can run on Gemini 3 Flash (default), Gemini 2.5 Flash or Pro, or GPT-5 / GPT-5 Mini. Pick per seat depending on speed vs depth.
How are tokens charged?
Creating a meeting, adding an AI seat, every AI reply, and generating a deliverable each cost tokens. Premium models cost a few more tokens per reply.
Do I need my own OpenAI or Google API key?
No. Tokens cover everything — model usage, hosting, storage. Bring-your-own-key is on the roadmap for enterprise plans.
Can I get a refund?
Unused token packs are refundable within 14 days of purchase. See the refunds page for details.
Privacy & data
Is my data private?
Every meeting is owner-scoped. Documents, transcripts, and tool calls are only visible to the host and invited members. We never train models on your content.
Can I export or delete my meetings?
Yes. Every meeting can be exported as Markdown or PDF, and deletion is permanent and immediate from the meeting settings.
Still have questions? Email us.
Run your first meeting with 500 free tokens
Sign up, cast the room, and stress-test your next decision in minutes.