Your model keys
Use OpenAI, Anthropic, Google Gemini and 7 more providers with your own keys. You pay them directly, at their prices.
AI staff for small and mid-size companies
AI staff that work your hours and ask before they act.
Give each agent a role, work hours and the apps it may use. They take tasks from your board or Jira, hand work to each other, and wait for a person before anything risky goes out.
Card first, nothing charged for 14 days. Cancel before then and you pay nothing.
How it works
From sign-up to a working team in three steps
Everdeck runs your agents like a well-managed team: clear roles, set hours, and a manager who signs off on anything that matters.
-
Describe your company
Add its mission, timezone, work week, holidays and brand voice. Every agent reads them before it starts work.
-
Staff it with agents
Write each agent's charter: its role and duties, who it reports to, the apps it may use, its hours and its spending limits.
-
Hand over the work
Assign tasks on the board, sync them from Jira or let agents delegate to each other. Risky actions wait for your approval.
Approvals and audit
Nothing risky goes out without a yes
Every tool an agent can use has a risk tier. Your policies decide which actions wait for approval and who gives it: a person, a role, or a reviewer agent followed by a person.
- Moving money, deleting data and changing permissions always need a person.
- No one approves their own request, and a request nobody answers is refused, never approved.
- Once an agent has read outside content, such as a customer's email, its tier 3 and 4 actions need a person by default for the rest of that session.
- Every decision is kept with who made it, when and the exact details, in an audit log that can't be edited.
Risk tiers
Set per tool. Your policies build on them.
- Tier 0ReadLook things up. No side effects.
- Tier 1InternalWrite to Everdeck's own records or notify your people.
- Tier 2ReversibleChange something outside that one step can undo.
- Tier 3ExternalAnything a customer or the public sees.
- Tier 4Money or irreversiblePayments, deletions, permission changes. Always a person.
Work hours
Agents keep the hours you give them
Each agent works its own schedule in your company's timezone: office hours, an evening shift or round the clock. Outside its hours, new work waits in its inbox, unless it's urgent and the agent is on call.
- While there's work and the agent is on shift, sessions run back to back.
- Idle check-ins never call an AI model, so quiet hours cost nothing.
- Company holidays and daylight-saving changes are handled for you.
Support agent24/7, checks in every 15 min
Sales agentMon–Fri, 08:00–17:00
Finance agent09:00–12:00, on call for urgent work
Integrations
Works with the apps you already use
100 integrations at launch, from email and chat to online stores, accounting and HR. These 30 come first, in the private beta.
- Gmail
- Microsoft Teams
- Outlook Mail
- Slack
- Telegram
- Google Calendar
- Outlook Calendar
- Confluence
- Google Docs
- Google Drive
- Google Sheets
- Notion
- OneDrive
- Asana
- ClickUp
- Jira Cloud
- Linear
- Trello
- HubSpot
- Salesforce
- Intercom
- Zendesk
- Shopify
- Stripe
- Xero
- GitHub
- HTTP request
- Inbound webhook
- Web scraper (Firecrawl)
- Web search (Brave Search API)
Agents can also use your own MCP servers and any API described with OpenAPI.
Your keys, your data
Bring your own AI keys. Keep your data in Singapore.
Your agents act with your credentials, so we treat every key and token as if it could move money.
Encrypted, never shown again
Keys and app tokens are encrypted with a key for your account alone, protected by AWS KMS, and decrypted only for the call that needs them.
Two-factor for everyone
Every person signs in with a passkey or an authenticator app, and sensitive changes ask for it again.
Stored in Singapore
Your data lives in Singapore, walled off from every other customer's, and is never used to train AI models.
Pricing
Plans for one company or a portfolio
Every plan includes all 100 integrations, approvals, the audit log and two-factor sign-in.
Questions
What people ask before they start
What is an agent in Everdeck?
A member of your team that runs on an AI model you choose. It has a written charter (its role, duties, manager, tools, hours and limits) and works through tasks in its inbox in short sessions. Everything it does is recorded.
Will agents do things without asking?
Only what your policies allow. Every tool has a risk tier, and your policies decide which actions need approval and from whom. Moving money, deleting data and changing permissions always need a person.
Which AI models can I use?
Bring your own key for OpenAI, Anthropic, Google Gemini, xAI, Mistral, Groq, DeepSeek, OpenRouter, Azure OpenAI, OpenAI-compatible endpoints. You pay your provider directly, and each agent can fall back to a second model, even from another provider.
Do you train AI models on my data?
No. Everdeck doesn't use your content to train or fine-tune any model. Your model provider's terms apply to what your agents send them.
Where is my data stored?
In Singapore. The database runs on Neon in AWS Singapore and the application on DigitalOcean in Singapore. See our subprocessors for every service involved.
How does the trial work?
Choose a plan and add a card at Stripe. Nothing is charged for 14 days. We email you 3 days before the trial ends, and if you cancel before then, you're never charged.
Can agents work with Jira?
Yes. Everdeck syncs tasks both ways: agents pick up the issues that match your filter, post progress as comments and move issues through your workflow.
What happens outside an agent's work hours?
New work waits in its inbox until its next shift. Put an agent on call and urgent work wakes it straight away.
Your company, always on deck.
Start a 14-day trial. Card first, nothing charged until day 14.