aramb./ edition 09
All systems nominalN° 09: Agent Economy
Vol. 09
Feature
The agent economy issue
Build AI agents
that do the work
you'drather not.
Describe it in plain English or write it in code, and Aramb runs it for you. Memory, voice, browser and 10,000+ tools come built in.Ship in an afternoon. Bill every user. Sleep on Sundays.
Your agent can now
book the meeting,
and bill for it.
chase the invoice,
politely.
triage the bug
before standup.
answer the customer
at 2am.
find tomorrow's prospects
while you sleep.
book the flight,
aisle seat, as usual.
answer the call,
on the first ring.
close the ticket,
before you see it.
run the report,
before Monday.
find the lead,
and warm it up.
screen the candidate,
before your first coffee.
draft the reply,
in your voice.
reconcile the books,
to the penny.
triage the inbox,
down to zero.
qualify the deal,
before the call.
onboard the customer,
without the kickoff call.
file the expense,
receipt and all.
book the meeting,
and bill for it.
Speaks fluently to ↓
OpenAIAnthropicGooglexAIGroqCartesiaElevenLabsDeepgramWhisperStripeGmailSlackLinearGitHubNotionOpenAIAnthropicGooglexAIGroqCartesiaElevenLabsDeepgramWhisperStripeGmailSlackLinearGitHubNotionOpenAIAnthropicGooglexAIGroqCartesiaElevenLabsDeepgramWhisperStripeGmailSlackLinearGitHubNotion
§ 01: Use cases
One platform.
Every kind of busywork.
Pick the job that eats your week. Each one starts from a working template.
For founders
Finds the right prospects each morning and drafts the first email in your voice.
Use this template→
For engineers
Triages new issues from GitHub and Linear and has a summary ready before standup.
Use this template→
For operations
Chases unpaid invoices, updates the ledger and flags the ones that need you.
Use this template→
For support
Answers customers on WhatsApp and email, and hands over when a person should step in.
Use this template→
For sales
Prepares you for every call and writes the follow up before you've closed the tab.
Use this template→
For you
Plans your week with you and checks in each day to keep you on track.
Use this template→
§ 02: Field guide
From idea to working agent
in three moves.
Are you a developer?
- 01DescribeTell Aramb what the agent should do and the rules it follows, in plain English. Or chat to Woky on WhatsApp.
- 02Try itTalk to it before anyone else does. Change the words, run it again.
- 03Switch it onOne click and it's live: on your site, on WhatsApp, on a schedule, or when something happens.
- 04Share it(optional)Built something good? Let others run it and earn on every use.
aramb · builder01 / 04 · Describe
Live in about 20 minutes · No servers · Idle agents cost nothing
§ 03: Start here
Not sure what to build?
Start with customers.
Our free growth agent finds the right prospects every day, drafts outreach in your voice and tells you what to do next. You approve everything before it goes out. Bring your own AI key, or connect Claude Code.
Get the free growth agent →Free · Your accounts, your approval
Stuck? A person from our team helps you set up your first agent.
- 01Around 20 prospects a day, each with a reason to reach out now
- 02Emails from your own inbox, LinkedIn notes for you to send
- 03A weekly review with your next best move
§ 04: Seven primitives
Seven primitives.
One control plane.
Every agent product converges on the same seven services. We built them so you don't stitch nine vendors and a Redis into a demo that dies on day two.
01
Agent
Any model. Any provider. Route on price, latency, or intent. Swap without a rewrite.
02
Voice
Sub-400ms realtime pipelines. Whisper, ElevenLabs, Cartesia, Deepgram, behind one interface.
03
Browser
Headless Chrome that survives captchas, logins, and 6-hour sessions. Every click is billed.
04
Sandbox
Ephemeral Linux boxes for tool-use. Fenced per end-user. Metered by the second.
05
Memory
Vector + episodic + semantic. One query surface. Namespaces per tenant, free of RAG plumbing.
06
Tools
1,500+ integrations pre-authed. OAuth, refresh, retries, rate-limits, solved once.
07
Billing
Every primitive emits a usage event. Pipe to Stripe, Paddle, or your ledger. No math.
§ 05: By the numbers
10,000+
Tools available
27ms
P50 tool latency
From $2.00
Per 1,000 credits on monthly plans
6 lines
To first agent
Built by people who got tired of gluing nine dashboards together to ship one agent. If that sounds familiar, you're the reason this exists.
the aramb team, writing from a warehouse in London
§ 06: Pricing
Pay for what
your agents actually do.
Start free. Every plan draws from a single pool of credits, with no per-seat fees. Add as many teammates as you like, and unlock unlimited workspaces and agents on paid plans.
Annual · 2 months free
Hobby
for prototypes
$0free forever
1,000 credits to start
then claim 1,000 every week
- — 1,000 credit sign-up bonus
- — Claim 1,000 credits every week
- — Bring your own model key*
- — Sandboxes included
Studio
for solo builders
$25/mo
11,000 credits/mo
$2.27 per 1,000 credits
- — Everything in Hobby, incl. the weekly claim
- — Dedicated compute
- — Unlimited workspaces, agents + seats
- — 30 min of expert video help a month
Scale
Most popular
$79/mo
39,000 credits/mo
$2.03 per 1,000 credits
- — Everything in Studio
- — Per-tenant metering: rebill your users
- — Usage analytics
- — 60 min of expert video help a month
Custom
for growing teams
$250/mo
Custom plan size: monthly credits
- 75,000 credits/mo$150/mo
- 100,000 credits/mo$200/mo
- 125,000 credits/mo$250/mo
- 175,000 credits/mo$350/mo
- 250,000 credits/mo$500/mo
- 375,000 credits/mo$750/mo
$2.00 per 1,000 credits
- — Everything in Scale
- — Our lowest rate on a monthly plan
- — Choose the monthly credits you need
- — Video help from an expert
* Bring your own model key: connect your own LLM API key and your provider bills you for the tokens directly. Aramb charges no overage for your first 3 months, then a 10% overage on the tokens used, charged in credits. Your credits also cover the rest of the platform: compute, browser, voice and tools.
// 1 credit covers it all: LLM tokens · CPU · memory · volumes · egress · captcha · proxies · integrations · voice.
Full pricing, top-ups & credit rates →§ 07: Margin notes
Small print,
big deal.
GDPR: in progressHIPAA: in progressSOC 2: in progress
Aramb Token Kompressor (ATK)
Compression that trims prompts and context to cut LLM token spend by up to 50%: same output, half the bill.
Self-learning by outcome
Agents improve from real results, not guesswork. Every outcome sharpens the next run.
Autonomous agent testing
Agents evaluate and regression-test themselves before anything reaches production.
Built-in agent audit
Every action is logged, traceable, and replayable, a full audit trail out of the box.
AI agent compliance
Built for regulated workloads. GDPR, HIPAA, and SOC 2 all in progress.
Metered per user
Every session is tagged to an end-user id. Pass it to Stripe or aggregate yourself. We don't care, we just emit.
No vendor lock
Swap Deepgram for Whisper by changing a string. Same interface, same bill, different weights.
Fenced sandboxes
Every code execution runs in an ephemeral VM. Nothing leaks between your customers, ever.
Region pinningComing soon
Route sessions to us-east, ldn, or fra. Data-residency by header, not by a support ticket.
Silent failoverComing soon
Providers go down. Your product doesn't. Auto-reroute across three vendors per primitive.
One invoice
Voice + browser + tools + models on a single line-itemed bill. Reconciliation, deleted.
§ 08: Compare
How aramb stacks up,
honestly.
Eight head-to-heads against the platforms you are probably already using, what each one is genuinely better at, and where we think we win.
Twin
vs aramb→
ChatGPT Agent
vs aramb→
Claude Cowork
vs aramb→
Dust
vs aramb→
Zapier
vs aramb→
Make
vs aramb→
Lindy
vs aramb→
Gumloop
vs aramb→
§ 09: Start building
Ship an agent
before your coffee
goes cold.
Get started →
Free to start. No card required.
