hurozo Attend
First step

Set up your first agency

Create an agency, add workers, attach canon, and assign its first task.

Worker Tools

Manage worker

Tune the role and tools for this agency teammate. Graph wiring stays internal.

Tools

Loading...

Worker View

No tools selected
Team worker Untitled Agent Describe what this agent should do. HurozoAgent

Worker Chat

Describe the worker you want to add to this team, or what you want to change.
Draft JSON 
{}

Teams

Teams are managed from Agency.

Connectors

Connectors

Connect services for the current agency

Loading connectors…

Agency

Team workspace

Planning

Filters

Build

Runs

Execution activity

Runs

Inspect Worker work, tool calls, reasoning summaries, and failures.

Select a run to inspect its activity.

Files

Agency files

    Documentation

    Contact Us

    New message

    To: hi@hurozo.com

    Hurozo CLI

    Terminal-first agent work

    Hurozo CLI

    The same Hurozo agent runtime, model routes, approvals, and connected tools from your shell.

    Install with Homebrew 
    brew tap hurozo/hurozo-cli
brew install hurozo-cli

    View latest GitHub release

    Hurozo CLI running in a terminal

    Model routing

    Switch between Hurozo, Gemini, Opus, and ChatGPT-backed routes when your account is configured for them.

    Approvals

    Review tool permissions from the terminal and keep remembered decisions scoped to the exact operation.

    Second opinions

    Use model arbitration to ask another available route for critique, checks, or a different perspective.

    Connected tools

    Use the same account-backed connectors and skills available in Hurozo's desktop and WebCLI experiences.

    First run

    After installation, run hurozo. The CLI walks you through account authentication and any model-specific sign-in or secret setup that your selected route requires.

    Hurozo SDK

    Python agent invocation

    Hurozo SDK

    Invoke saved Hurozo agents from Python using the same API tokens, inputs, outputs, and runtime routing used by Agent Builder.

    Install 
    pip install hurozo
    getting_started.pyPython
    import json
import os

from hurozo import Agent

os.environ["HUROZO_API_TOKEN"] = "YOUR_API_TOKEN"
agent = Agent("sentiment-analyser")
inputs = {
    "prompt": "Analyze sentiment and return JSON. Text:",
    "input": "I am very happy!",
}

result = agent.run(inputs)
print(json.dumps(result, indent=2))

    Agent lookup

    Use an agent name, saved agent id, or agent UUID. The SDK resolves the correct execution target.

    Structured inputs

    Pass a Python dictionary matching the API input keys exposed by the saved workflow.

    Hosted runtime

    Runs against Hurozo's execution endpoint, including model routing, tools, outputs, and usage accounting.

    Token setup

    Create an API token from Account settings -> Manage API Tokens, then set it as HUROZO_API_TOKEN before running scripts.

    How you get started

    Prove one process delivers, before you commit.

    A guided sprint on your own tickets. We set up a governed crew, prove it on one process against your baseline, and hand you the keys. You own what you build.

    Book a demo See it work
    The sprint

    A mutual action plan, milestone by milestone.

    Not a fixed number of days. A short proof up front, then a longer embed, each phase with an output you can see.

    Milestone
    First result
    Day 1. The crew drafts on your real backlog, held at the gate.
    Milestone
    Proof
    ~2 weeks. Enough observe-mode runs on one process to trust it.
    Milestone
    Embedded
    ~4 to 6 weeks. Governed go-live, keys yours, expansion underway.
    0
    Scope
    Agree the pilot from your own AI-ify list, the list of what you want to transform. One process, chosen with you, not the whole desk.
    the signed mutual action plan
    1
    Baseline
    Connect your first assets and databases. Capture how support runs today via recorded interviews with your team, run by our voice agent. We set the number to beat.
    your baseline, plus connected sources
    2
    Ingest and canon
    Your policies and assets ingested and checked for collisions, gaps, and opportunities. The crew generates your first canon version, reviewed with your point person.
    your first canon, plus a gap and opportunity map
    3
    Proof the trust ladder
    No big-bang cutover. One process climbs rung by rung, each earned on evidence, measured against your baseline.
    1. Replay. Run the crew over your last few hundred closed tickets. Proof on real data, zero live risk.
    2. Mirror. Live shadow, draft-only, no write access to your systems. Track the agreement scorecard: the crew's call against your team's.
    3. Canary. First live actions on a thin slice: one queue, low-risk intents. Blast radius contained.
    4. Assist. Act-with-approval on the proven category. You hold the gate.
    5. Autonomy. Auto for a category only once it clears the bar. Everything else still escalates.
    the agreement scorecard, plus the proven process
    4
    Run
    Governed go-live on the proven process, act-with-approval. The keys are handed over. The process runs, and it is yours.
    one process live and owned
    5
    Expand
    Work the expansion plan: the next processes, the next crews, coverage climbing. The pilot is a first step, not a one-off.
    the expansion plan in motion

    Day 1, ~2 weeks, and ~4 to 6 weeks are how we design the rollout, calibrated on the first pilots, not promised results. The drafts are real, on your data, in observe mode. Nothing is sent until you approve it.

    The proof, on your data

    Every action drafted, logged, and held for you.

    Attend · your support deskobserve mode · sample data
    Where is my order #4471routine
    Refund, arrived damagedneeds reply
    Reply drafted from your refund policy · your call
    Read it firstApprove and send
    permissiondraftheld at gatemodel loggedgrounded in canon
    Agreement scorecardmirror rung · sample
    TicketCrew wouldYour team didVerdict
    Where is order #4471Share trackingShare trackingagree
    Refund, arrived damagedRefund and apologiseRefund and apologiseagree
    Cancel, changed mindCancel and refundOffer to pause firstdiverged
    Each divergence becomes a canon update or a labelled example. The scorecard is the artifact your champion carries to the exec, and it seeds your canon at the same time.

    This is the Mirror rung: live shadow, draft-only, no write access to your systems. Observe mode on sample data here. On your own tickets, nothing is sent until you approve it, and every action is measured against your baseline.

    What it takes

    What you bring, what we bring.

    You bring
    The inputs only you have
    • 01Your policy and help-center docs.
    • 02Access to your help desk, CRM, or mailbox.
    • 03A point person for the sprint.
    • 04The approvals. You hold the gate.
    We bring
    Everything else
    • ✓The crew, pre-built.
    • ✓Your starter canon, drafted from your docs.
    • ✓The connectors and setup.
    • ✓The gate, roles, and the audit trail.
    • ✓The runbook, and a person for the custom calls.

    That is the whole ask. If your docs and your help desk are ready, so are we.

    What you keep

    You own what you build. It compounds.

    ✓
    A proven process
    ✓
    Your baseline captured
    ✓
    Your first canon
    ✓
    A running governed crew
    ✓
    The expansion plan
    Governed resolution ratesample
    Tickets resolved on-policy and provably. The number comes from your run, not from us.
    Canon coveragesample
    Answered from your own canon, not a guess. Climbs the longer you run.

    Your canon and your audit trail live in your workspace, and the audit trail supports the record-keeping the EU AI Act calls for. Not lock-in. Leaving means leaving your own work behind.

    How we run it

    Mutual, milestone-gated, and observed before it acts.

    A mutual action plan
    Both sides sign the plan in Phase 0: the scope, the goals, the success criteria, the people, the dated milestones. Nothing waits on an unnamed owner.
    Run by the onboarding crew
    Most of the setup and the first canon are drafted by our onboarding crew, the same kind you are about to operate. People step in only for the custom calls.
    Milestone-gated
    A short proof up front moves the decision, then a longer embed earns the trust. You see an output at every phase, not just at the end.
    Observe before act
    The crew observes on real tickets first, drafts held at the gate. It goes live only on the proven process, and only once your approver says so.
    Book a demo See it work