Free
Messaging and a home for your own AI
$0
Get started for freeKey features:
- Unlimited people and messages
- Last 90 days of history visible
- Chat, tools and Loops on your own model keys
- AI spend dashboard
- 1 workspace, 1 GB wiki
Free to talk. Pay for the agent.
Free as a messaging app and a home for your own AI models. Pro and Scale add the Harmony Agent, billed per active user.
Annual billing selected. Prices shown per month, billed annually.
Most popular
Pro
The Harmony Agent and the whole product
$16.67 /user/month
Get startedEverything in Free, plus:
- Harmony Agent with standard usage limits
- Unlimited history, Syncs with recording and summaries
- Background agents, Loops and model routing
- 5 workspaces, 100 GB wiki with daily sync
- Lifecycle Graph and Perception
- Business-hours support
Scale
For heavy agent users, alongside Pro
$82.5 /user/month
Get startedEverything in Pro, plus:
- 10× Pro’s Harmony Agent usage
- Cloud computer hours included
- Budgets on AI spend by team and person
- Unlimited workspaces, 500 GB wiki with hourly sync
- Audit log without a time limit
- 24/7 support
Enterprise
An organization-wide contract
Custom
Contact usPro and Scale seats, plus:
- Pooled usage set in the contract
- Single sign-on (SAML and SCIM)
- Executive AI spend dashboard
- Dedicated region, data residency and customer-managed keys
- Named contact
Pricing questions.
Who is billed?
An organization starts free and upgrades as a whole. Every active user then pays at least Pro, and admins move heavy users to Scale. Members who were invited but don't use Harmony are never billed.
How do Harmony Agent usage limits work?
Each person's Harmony Agent usage has a five-hour limit and a weekly limit. You see the share of each limit you've used and when it resets, with a warning at 80%. Scale gives ten times Pro's usage in both. Limits are per person, so one heavy user can't use up the team's allowance; Enterprise contracts can pool usage.
What counts against the limit?
Agent chat, tool calls, Loops, background agents, the agent in Syncs, reports and wiki indexing, whenever they run on Harmony's models. Anything that runs on your own model keys doesn't count, and work that fails or is cancelled never counts.
What happens at the limit?
Three choices. Wait for the reset, buy extra usage at model cost plus 30% (admins set a monthly spending cap), or switch to your own model key.
Can we use our own AI models?
Yes, on every plan including Free. Bring keys for Anthropic, OpenAI, Gemini, xAI, Mistral or OpenRouter. Your provider bills the model usage and Harmony adds 10% on the AI spend it can see, including declared subscriptions and other AI tools on the spend dashboard. Invoices under £5 a month are waived.
What are add-ons?
Extra usage, wiki storage, extra workspaces, Sync transcription hours, cloud computer hours, premium data sources, real-time connector sync and API and MCP call packs, billed monthly with your plan. Usage-based add-ons stop at the spending cap an admin sets. Free can buy API and MCP call packs.
What does Enterprise add?
Enterprise is an annual contract on top of Pro and Scale seats, with pooled usage, single sign-on (SAML and SCIM), the executive AI spend dashboard, a dedicated region, data residency and customer-managed keys.
Can I try the Harmony Agent before paying?
Yes. Everyone on Free gets a small one-off allowance of Harmony Agent usage, so you can see what it does before you upgrade.
Get started for free.
Bring your team and your own AI models. Add the Harmony Agent when you're ready.