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.

    Finally.

    Automate customer service true to your brand.

    It answers every customer from your own truth, on every channel, and hands the hard ones to a person.

    The same answer, in your voice, whether they call, email, or message. Never four bots drifting apart.
    Book a demo See it work
    Your keys · every answer logged · yours to keep
    From your truth · your voice · every channel
    See it work

    Watch it run a whole morning.

    The whole desk in motion, one ticket to the gate, the one thing to fix, the canon it all runs on, what it costs, and the numbers you would show your board. One desk you own.

    Attend · Your support desklive · sample data
    The 9am wall
    Read and sorted overnight. You walk in caught up.
    294 handled · 18 held
    The ticket journey
    One ticket, intake to resolved. It stops at the gate, only when a person should decide.
    the record
    What to fix
    The desk reads every conversation and names the one thing to change.
    this month
    The issues that came back most this month, by tickets
    Refund-window confusion47
    Damaged on arrival31
    Wrong size19
    Fix the top oneRefund-window confusion is your biggest recurring cost, 47 tickets this month. State the window as 14 days, up front. Approve once, it lands on every channel, and those 47 tickets stop.
    The canon
    The governed corpus every answer draws from, and how healthy it is.
    live
    96%
    answers on-canon
    92%
    coverage
    1,338
    sources
    3
    flagged for you
    Coverage, last 90 days61% → 92%
    Resolved tickets1,050
    Product docs210
    Your policies40
    Your calls at the gate+38
    Every answer is grounded and logged. Drift on any surface, off-fact or off-voice, gets flagged for your approval before it spreads.
    Cost & models
    Your keys, your models, at cost. The router picks the cheapest that clears the bar.
    this month
    €0.04 / answer
    down from €0.09 in March
    93% of contacts never reached your team. No markup on any answer, and no per-resolution tax on your own customers.
    Where this month's work ran
    Claude62%
    Gemini31%
    GPT7%
    Your keys, your model. Swap models without touching a workflow.
    The Monday report
    Yesterday, by case type and how much a person had to touch.
    yesterday
    312
    cases
    84%
    resolved on their own
    12%
    held for you
    4%
    to a person
    92
    64
    48
    39
    34
    21
    14
    OrdersDeliveryReturnsRefundProductBillingComplaint
    Resolved on its ownHeld for youTo a person
    Refunds and complaints are where your people are needed. Everything routine answered itself, grounded and logged.
    One desk, one canon, one gate. Owned and yours to keep.

    Six views here. On a call, we run yours.

    Own it. Prove it. Never a wall.

    Run the queue, escalate the hard calls.

    One crew runs the whole desk. Here is each job on its own. Open any one to see it work.

    The morning wall

    Every message read and sorted overnight, with the one that needs you on top.
    Where is my order? order
    Change my address account
    Damaged on arrival returns
    Angry about a delay needs you
    See it work →

    The reply, already written

    Drafted in your voice, ready before you open the ticket.
    Draft · in your voice Hi Sanne, good news, your order shipped this morning and is due Thursday. Here is the tracking
    See it work →

    The right answer, every time

    From your own policies and product truth, never the open web.
    Yes, sizes run true. This jacket is a regular fit, so take your usual size.
    ✓from your canon + Odoo
    See it work →

    Held for your yes

    Refunds and one-off calls wait for your approval, never sent alone.
    Held for your call
    Refund €80, 10 days past window. Loyal customer, 6 orders.
    ApproveChangeDecline
    See it work →

    Straight to a person

    The upset or tricky one reaches a teammate, in time.
    M
    To a personSecond complaint this week. Handed to Marta with the full thread.
    See it work →

    A record you can show

    Every answer logged and inspectable, for leadership or the regulator.
    ✓answered from canon#A-4471
    ✓held, approved by you#A-4472
    Art. 12 logsArt. 14 oversight
    See it work →
    ▣
    See the whole loop

    One ticket, five stations, your call in the middle.

    Watch the desk run →
    The problem

    Where customer service breaks, every day.

    You are firefighting the queue by hand. The team is stressed, and nothing is getting better. You know you should automate it. You just have not found a way you would put your brand on.

    A normal morning
    40 unread by 9am, and climbingno time
    the same question, for the fortieth timeagain
    the angry one, missed until it is publictoo late
    a new hire guessing at the refund policyoff script

    "We keep saying we should automate it. Nobody knows how to do it in a way I would trust."

    Head of customer service · sample

    "We rigged up something with ChatGPT and n8n. It went rogue, and I could not put our name on it."

    Same person, three months later · sample
    Everyone says automate. Nobody hands you an AI you would actually trust with your customers, and your name.
    Interoperable

    It plugs into what you already run.

    Answers from your live systems and lands on the channels your customers already use. It leverages what you have, no rip and replace.

    WhatsAppEmailPhoneWeb chatOdooYour CRMJira
    See integrations
    Secure & governed

    Built to pass your security review.

    Your own keys, permissions per worker, and Worker runs recorded as they happen, designed around the EU AI Act. EU residency is on our roadmap.

    EU residency: roadmapYour keysRoles + gatesRun logsEU AI Act, by design
    For IT & security
    Try it yourself, right now.
    Message our own agent on WhatsApp right now and watch it answer, grounded and gated.
    Message the agent now
    Pricing

    Pay for what it runs, not per seat.

    You bring your own model keys and pay providers at cost. No markup on every answer, no rising tax on your own customers.

    Solo
    Free
    One worker on your inbox, watch-only. The activation moment.
    Crew
    from €490 / mo
    The full support crew, live, gate on every send, your canon and audit trail.
    Most brands start here
    Department
    €1,900 / mo
    The cockpit: analytics, the insight loop, SSO and roles, higher volume.
    Workforce
    from €7,500 / mo
    Multiple departments and a named SLA. Self-hosting and EU residency on the roadmap.

    Draft pricing, owner-gated. Final numbers are set with you on the call.

    See full pricing →
    Be known for your service

    Scale the service you're proud of.

    The crew answers on every channel, from your own truth, and holds the calls that need you. You own it, and you can prove it.

    Book a demo See it work →
    You own it Provable, shaped to the EU AI Act EU residency on our roadmap Never a wall
    Attendlive
    Can I return this jacket? Bought it 40 days ago.
    Our window is 30 days, so this one is a judgment call. I have asked a colleague to take a look.
    Held for a person
    Marta from the team is on it. Out-of-policy return, your call.
    Real Attend surface, sample data. Not a mockup.