Cortexly
Create your cortex
FeaturesSix faculties of the cortexDemoWatch one digest travel end to endPricingOne balance — nothing renewsFAQBefore you hand over a routine
pay per run, never per seatCreate your cortex
The operations cortex

A second brain for the work that repeats.

Cortexly keeps your recurring operations — digests, reports, monitoring, follow-ups — as crews of AI agents that run on schedule, hold their context and settle against one token balance.

Create your cortexWatch a routine run
ledger — append-only
routine reserve− 60013,780
run charge · r_9d27− 52613,854
hold released+ 7413,854
every line remembers the balance after
model mix — per agent
claude-sonnet-5gpt-5.2gemini-3.5-flashgrok-4

each mind in a routine picks the reasoning its step needs

No subscriptions, ever330+ ready-made crewsBring your own model keys
0+crew templates you can adopt as-is
0+connectors, from Slack to raw HTTP
0%of debits recorded with the balance after
Nomonthly plans — you top up, crews spend
Capabilities

What the cortex holds for you

Six faculties, each doing one job well. Together they carry the routine load your team shouldn't.

Routines as living graphs

Sketch agents and hand-offs on a canvas. The graph is the runbook — and the runbook executes itself.

Runs on your rhythm

Trigger crews on a schedule, from a webhook or by hand. The cortex never forgets a Tuesday report.

Reach into 120+ services

Slack, Telegram, RSS, webhooks, plain HTTP — each agent gets exactly the access its step requires.

Model choice per agent

Attach your own Anthropic, OpenAI, Gemini or xAI keys and give every agent the reasoning that suits its task.

A memory of every token

The ledger is append-only: each debit lands with the balance after it, so cost questions have one answer.

Recall over REST

Any crew is one authenticated POST away. External systems start runs and collect results through the public API.

Watch it think

One routine, end to end

Follow a weekly competitive digest from canvas to Slack — including the exact ledger line it leaves behind.

S
Scanner

Sweep the rival watchlist

gemini-3.5-flash
A
Analyst

Distill what changed and why

claude-sonnet-5
E
Editor

Shape the digest for Monday

gpt-5.2
A scanner, an analyst and an editor each receive a role, a goal and a model of their own.
Pricing

One balance. Every routine draws from it.

Pick a pack size, top up, and let crews spend as they run. Larger packs lower the per-token rate — and nothing about it recurs.

≈ 2 routine runs
  • All 6 model providers
  • Sequential process
  • Built-in tools & actions
Top up Starter
≈ 5 routine runs+100 bonus tokens on this top-up
  • Hierarchical orchestration
  • Public API + 3 keys
  • Crew memory & templates
Top up Maker
≈ 10 routine runs+300 bonus tokens on this top-up
  • Everything in Maker
  • Priority queue lane
  • Scheduled triggers
Top up Pro
≈ 27 routine runs+1,000 bonus tokens on this top-up
  • Unlimited API keys
  • Priority queue lane
  • Low-balance alerts
Top up Scale
≈ 60 routine runs+3,000 bonus tokens on this top-up
  • Everything in Scale
  • Team invites
  • Best per-token rate
Top up Studio

Tokens never expire and there is nothing to cancel.

FAQ

Before you hand over a routine

recall 01 / 06

What does a power token actually pay for?

Model calls, connector actions and engine time all convert into one unit. Your balance is a single integer, and every run draws it down transparently.

focus here

Move your first routine in today.

Sign up, claim your starting tokens and hand over one recurring task before lunch.

Create your cortex