ProductAlways-on AI assistant
PricingUNLI- use · flat monthly
StatusService online

Hermes Agent, configured for you in one click.

UNLI- prompts. UNLI- replies. No token caps on gemma4:26b.

No VPS to manage. No token meter to watch. Always on. Pre-configured agent running on gemma4:26b — you tell it about your business, it goes to work. ₱2,999/month, cancel anytime.

The Alagad — armored guardian in salakot, eight-rayed sun of the Philippine flag with the Baybayin a on the chestplate
Live inference · serving right now
Live gemma4:26b inference
~66tok/s
Decode speed · 24h average
All systems operational · 2 GPU clusters serving 30-day uptime 99.7% · updated 12m ago

Freedom from token limits.

The AI is included. One flat monthly rate, no metering, no top-ups, no surprise bills. Use your assistant the way you'd use any other tool you bought outright — until it gets the job done, not until your credits run out.

Talks to your customers

Channels

Your assistant lives on WhatsApp, Messenger, Telegram, Discord, and the other apps your customers already message you on. They reach the AI the same way they reach you today.

MessagingWhatsApp · Messenger · Telegram Also onDiscord · Slack · 15+ more VoiceVoice notes on Telegram & Discord LanguagesEnglish · Tagalog · Taglish

Remembers every conversation

Memory

The assistant remembers what customers told you last week, last month, and last quarter. It builds a profile per customer and carries context across every chat — so nobody has to re-explain themselves.

Customer notesper-contact memory Conversation historyunlimited, searchable Business knowledgeyou teach it once, it sticks

Open AI model from Google's Gemma family

UNLI-

The AI behind your assistant is gemma4:26b, a mixture-of-experts (MOE) open-weights model from Google's Gemma family. MOE means more tokens-per-second per peso of GPU — which is exactly what makes flat-rate pricing work. We run it on our own hardware so you don't pay per message.

Modelgemma4:26b FamilyGemma · MOE · open weights Pricingincluded · UNLI- Usageno meter, no cap

Learns your way of working

Skills

Teach it once how you handle bookings, quote requests, or follow-ups, and it uses that pattern every time after. The longer it works for you, the sharper it gets.

Custom skillsadd via plain text Scheduled tasksmorning briefings, follow-ups Tool accessread your docs, fetch URLs Updatesyour prompts stay yours

No credits to top up. No meter to watch.

Most AI tools charge you per message. We don't — because we run the AI ourselves. Your subscription pays for the assistant; you don't pay again every time a customer chats. No daily request caps. No "fast tier" that throttles to a "slow tier." Same gemma4:26b on every request, at the same speed, all month.

What's included · gemma4:26b · Gemma family · MOE Live · serving
The AI
gemma4:26b · Gemma · MOE
Memory
across every conversation
Messaging channels
WhatsApp · Messenger · more
Voice mode
send & receive voice notes
Pricing
flat · UNLI-
Add-ons
none — everything's in
Backup AI
warm secondary
Messages per month
Use it like a staff member, not a meter. Sized so every customer can chat at peak hours without slowing down.
0
Per-message charges
Your subscription buys the assistant outright. No credit balance, nothing to top up, nothing to run out of.
15
Founding spots
Capped to keep response times fast. The next cohort opens with the next batch of capacity.
UNLI-

If you're Filipino, you already know how this works.

Unli rice. Unli text. Unli load. Pay once, use it like the meter doesn't exist — because it doesn't. We brought that to AI. UNLI-replies. UNLI-conversations. UNLI-everything. Whatever your customers throw at the assistant, the bill stays the same.

What your assistant can actually do.

One assistant, on call around the clock, doing the things you'd rather not handle at 11pm.

💬

Answer customer questions

Reply to FAQs, hours, pricing, and product details — in your tone, with your real knowledge.

📦

Take orders & bookings

Walk customers through ordering or booking, confirm details, and pass the order to you cleanly.

🔔

Follow up automatically

Send reminders, check in on quotes, and nudge no-replies — without you remembering to.

🗓️

Manage schedules

Set appointments, send confirmations, and reschedule when customers can't make it.

📊

Daily & weekly digests

Wake up to a summary of yesterday's customer chats, open questions, and anything that needs you.

🌐

Reply in any language

English, Tagalog, Taglish, Cebuano — your assistant adapts to how each customer writes to you.

🧾

Handle invoices & receipts

Send receipts, chase unpaid invoices, and answer billing questions before they reach you.

🛍️

Recommend products

Help customers pick the right item from your catalog, with the context of what they bought before.

🛠️

Whatever you teach it

Show it once how you handle something specific, and it uses that pattern from then on.

You run your business. We run the AI underneath.

Alagad gives you the assistant and the AI to run it — set up and ready to go. You decide how it sounds, what it does, and which customers it talks to. We handle the boring infrastructure parts.

[YOU]

You handle

  • Telling the assistant about your business
  • Connecting it to your messaging accounts
  • Deciding the tone and what it can answer
  • Knowing your customers — the assistant learns from you
[US]

We handle

  • Setting up your assistant and getting it ready to talk to customers
  • Running the AI and keeping it warm and fast
  • The hosting, the network, the GPUs, the updates
  • If anything underneath breaks, we fix it

One plan. One price. No tricks.

No tiers to compare, no promo-then-renewal trap, no setup fee. Same assistant, same UNLI- AI, same price — every month, for everyone.

One plan · month-to-month · cancel anytime Alagad Standard

Alagad Standard

  • AI assistant set up and ready to talk to customers
  • UNLI- messages · powered by gemma4:26b (Gemma family · MOE)
  • WhatsApp, Messenger, Telegram, and 15+ more channels
  • Memory across every conversation, per customer
  • Self-service dashboard — see usage, manage skills, restart anytime
  • Voice notes on Telegram and Discord
No contracts. Month-to-month billing. Cancel any time and the service stops at the end of the period. Uptime commitment: 99.5%. If we cause more than 4 hours of downtime in a calendar month, your next month is prorated as credit. No setup fee. No hidden token fees. The price you see is the price you pay on day one and day 365.
2,999 per month · flat
Billed in pesos. Local invoice. No USD card needed.

A fair deal, plainly stated.

Alagad gives you unlimited messages because we believe customers shouldn't get billed for using the thing they paid for. In return, we ask that you use it for a real business, within reasonable bounds, on the right side of the law.

For your own business

Run your assistant for your own shop, your own customers, your own work. The assistant and the AI are yours to use however your business needs them.

Not as a wholesale AI provider

Please don't repackage our AI and sell it as your own AI API to other companies. Building products on top of Alagad and selling those products is fine. Reselling raw AI access is not.

Not for training AI models

The AI is here to run your assistant, not to train a new model. Training, fine-tuning, and large-scale synthetic-data generation use the GPUs very differently — the flat rate isn't priced for that.

Within reasonable concurrency

Each assistant can handle 3 simultaneous chats. That's plenty for a real-world workload. If your workload grows past that, we'd rather size you correctly than have you fighting the queue.

Legitimate businesses only

We host serious businesses running legal, legitimate operations. Alagad is not available for, and we will not knowingly host:

  • Illegal activity under Philippine or U.S. law
  • Phishing, fraud, deceptive impersonation
  • Malware, ransomware, or unauthorized intrusion
  • Child sexual abuse material — zero tolerance, reported to authorities
  • Online gambling, lotteries, or sweepstakes without a valid license
+ 7 more categories
  • Spam, unsolicited bulk messaging, or unconsented outreach at scale
  • Adult, sexually explicit, or pornographic content of any kind
  • Ponzi, pyramid, HYIP, MLM, or make-money-fast schemes
  • Unlicensed pharmacy or controlled-substance distribution
  • Content inciting violence, terrorism, hate, or targeted harassment
  • Cryptocurrency mining or other workloads designed to abuse shared compute
  • Infringement of intellectual property, trademark, or copyright
Violation = termination. No refund, no notice beyond what the law requires. We cooperate with valid law-enforcement requests.

Common questions.

Do I need to know anything technical to use this?
+
The basics, yes — or someone helping you who does. The assistant is set up and ready to go, but you'll need to connect your messaging accounts (WhatsApp, Messenger, etc.) and tell the assistant about your business. We'll walk you through it. If you'd rather have someone else handle the setup, we can match you with an implementation partner.
What does UNLI- actually mean?
+
Just like unli rice means you eat as much rice as you want for one price — UNLI- means your assistant talks to as many customers as your business needs, for one flat monthly rate. No per-message charges, no credits to top up, no rate limits. The bill is the same whether your assistant handles 10 chats a month or 10,000.
Which messaging apps does it work on?
+
WhatsApp, Messenger, Telegram, Discord, Instagram DMs, Slack, SMS, email, and over a dozen more. The assistant is one brain across all of them — a customer can switch from WhatsApp to Messenger mid-conversation and pick up where they left off.
Does the assistant really remember customers?
+
Yes. It builds a profile for each customer — their preferences, past orders, what you told them last time — and carries that context into every future chat. It also remembers things you teach it about your business, your tone, and the way you like things handled.
Why gemma4:26b specifically?
+
It's a mixture-of-experts (MOE) model from Google's Gemma family, with open weights — which means we can run it on our own hardware without paying anyone per message. MOE matters because it routes each token through only a fraction of the parameters, so you get higher throughput per peso of GPU than a dense model the same size. That's what makes UNLI- pricing actually work. Sticking with one model also keeps your assistant's behavior consistent over time.
Is my data private?
+
Your assistant has its own isolated runtime, separate from every other tenant. Your customer conversations, your business knowledge, and your memory files stay yours. We don't read your chats and we don't train any models on them.
Who fixes what when something breaks?
+
We fix the infrastructure underneath — the runtime, the network, the AI itself. If something there breaks, that's on us. If your assistant is giving the wrong answers because of how it was set up, that's something you (or your implementation partner) can adjust in the prompt. We're around to help either way.
Can I cancel anytime?
+
Yes. No contract, no minimum, no cancellation fee. You can cancel from the dashboard any time and the service ends at the close of the billing period. Your slot opens up for the waitlist.

Get started.

Tell us about your business and what you'd want the assistant to do first. We respond within 48 hours and let you know whether you're a fit for the founding cohort.

Once you're in: your assistant gets set up and ready to go, credentials handed over, and we walk you through connecting your channels.

Response window ≤ 48h
Setup time same day after approval
Billing monthly · cancel anytime
Founding-cohort spots 15 maximum
Founding · application ● Open
Almost there. Your email client should have opened with your application pre-filled. Just hit send — it goes to [email protected] and we reply within 48 hours.