Full message history on every plan, including Free. Pay for seats; connecting your own Claude, ChatGPT or Cursor never costs extra.
Small teams trying it out
Growing teams that run on chat
Companies that need governance and AI depth
Dedicated infrastructure and contracts
Private deployment in your own cloud or data center
Free on every plan, no usage limits beyond standard rate limiting. It uses your seat — your permissions, your audit trail.
Agents with their own account (support desk, on-call bot, a partner’s SaaS agent). Free plan: 1. Paid plans: 1 included per 10 paid seats, then $5 / agent / month.
$10 per 1,000 credits, 10% off when bought annually. Credits are pooled across the workspace.
One credit is roughly one standard AI action — for example summarizing about 50 messages or translating one screen of text. Lighter tasks run on a fast model and cost 1 credit; heavier tasks that need a larger model (like drafting a quarterly review) cost about 6.
Credits are pooled across the whole workspace, so heavy and light users balance out. When the pool runs out, built-in AI falls back to a smaller model instead of stopping, and admins can set a monthly cap. The admin console shows usage by person, feature and model.
No. When you connect your own AI through MCP or our OpenAPI, the model runs on your AI provider’s account, so it never uses Zeze credits and is free on every plan. Credits only apply to AI features built into Zeze.
Bring your own key: on paid plans you can plug in your own Anthropic, OpenAI, AWS Bedrock, Google Vertex or any OpenAI-compatible endpoint — including a self-hosted model — for built-in AI features. Calls made with your key don’t use credits and carry no platform fee. Useful if you have committed spend with a provider, or need a specific model for compliance.
BYOK is rolling out together with built-in AI features.
A seat is a person who can sign in to your workspace. Deactivated members don’t count.
People from a partner company you collaborate with belong to their own workspace, so they don’t take seats in yours.
An AI acting on someone’s behalf uses that person’s seat. Standalone AI members are billed separately (see above).
Because losing old messages is the worst reason to switch tools. Free is limited by team size (20 people) and storage (10 GB), not by how far back you can scroll or search.
You run the full stack — messaging server, gateway, search, meetings — in your own cloud account or data center, using our deployment package. You bring your own AI models; none of your data has to touch our infrastructure.
Pricing starts at $22 per user per month billed annually, with a minimum of 250 seats or $50,000 a year, plus a one-time implementation fee. It includes updates, long-term-support releases and tier-2 support. If you’d rather we operate it inside your cloud account, ask about managed-in-your-cloud on the Enterprise plan.
Team and Business can be paid monthly at a higher rate ($10 and $19 per user). Enterprise and private deployment are annual contracts.
Talk to us if you are a non-profit, an education institution or an early-stage startup.