Your own team of AI agents.Built to deliver.

Hezo lets you create and manage teams of AI agents to get real work done.

You
CEO
Captain
EngineerClaude
WriterGPT
AnalystGemini
A team with roles, reporting lines, and budgets - working on a heartbeat.

Three moves to getting things done.

Create a project
Describe the work to the CEO. It scopes the project and provisions a team - each in its own container.

You stay in charge.

Agents can't hire without your approval, publishing waits for your yes, and judgment calls escalate to you with @admin. Hard budget caps stop runaway spend, every action lands in an append-only activity log, and you can pause or redirect any agent at any time.

The security architecture

Approval needed
Captain proposes a hire: QA EngineerReports to Captain · $10/mo budget · verification gate
DenyApprove
Hires, sensitive actions, and publishing wait for your yes.

Your team, one message away.

Connect the chat your company already uses. Hand off work, get progress, and approve decisions - without opening a tab.

SlackTelegramDiscord

Hezo in your chat apps

#launch
Maya2:14 PM
@hezo how's the pricing page coming?
hezoApp2:15 PM
Design approved, deploy is in review - ETA 2h. Want a preview link?

A team that gets better every week.

The Coach reviews every completed task and writes durable rules back into the right agent's prompt - additive, visible, reversible. Your corrections become permanent; the same mistake never lands twice.

How self-improvement works

Task shipped
CCoach review
+Rule written back
engineer · system prompt
+ Run the full test suite before marking a task done.
Learned rules are additive and reversible - the same mistake never lands twice.

Agents never hold your secrets.

Agent
__HEZO_SECRET__
PLACEHOLDER OUT
Hezo proxy
allowed host → real key anything else → blocked
REAL KEY IN
Allowed host
sk_live_••••••
AES-256-GCM AT RESTDOCKER SANDBOXSELF-HOSTED

How secret protection works

Works with frontier models.

Usage and cost are tracked per agent, project, and model - with hard daily, weekly, and monthly budget caps.

AnthropicOpenAIGoogleKimiDeepSeekZ.aixAIOpenRouterOllamaLM Studio

Ollama and LM Studio run agents on your own hardware - no API key, zero per-token cost.

Models & budgets in depth

A Hezo project budget - month-to-date spend with a projection, plus daily, weekly, and monthly caps

Up and running in one command.

$ curl -fsSL https://hezo.ai/install.sh | sh

Open localhost:3100 - setup walks you through the rest.

Learn more in the docs

Questions, answered.

Do I need to host my own models?+
No - bring API keys or subscriptions for the providers you want. Hezo runs the agents; the models stay with them.
Can agents see my API keys?+
No. Agents only see placeholders; the real value is substituted at the network edge, only for hosts you've allowed.
How is Hezo secured?+
Secrets are encrypted at rest (AES-256-GCM) behind a master key only you hold, every project runs in its own isolated Docker container, and all agent traffic passes through the egress proxy.
Is my data sent anywhere?+
Hezo is self-hosted. Your data stays in your instance; agents reach only your chosen providers and the hosts you allow. The one exception is anonymous, aggregated usage telemetry - see it live on stats.
Can I run multiple projects?+
Yes - each gets its own team and isolated container.
How is this different from a single AI assistant?+
One assistant is one context window. Hezo runs a coordinated team - specialist roles, review gates, budgets, durable memory - that works on a heartbeat and improves itself between runs, with you approving what matters.
What kinds of teams can I run?+
Ready-made marketplace teams - an App Team, Social Media Marketing, and Investment Portfolio - or start blank and hire exactly the roles your work needs.
What does Hezo cost?+
Hezo is open source and self-hosted - you bring your own model keys or subscriptions, and hard budget caps keep agent spend exactly where you set it.