Skip to main content

Automations

Chunk works while you sleep.

Research that re-runs itself. Watchers on any page. Agents that draft while you're away. Digests of your own notes. Set them once — Chunk runs the night shift and only pings you when something real happened.

Runs while you sleepQuiet unless it mattersEverything filed for you

☾ TONIGHT · WHILE YOU SLEEP

2:14 AM · RESEARCH

Solid-state batteries

3 new findings, cited

3:02 AM · WATCHER

Flight LHR→NRT

dropped $118 below your number

4:26 AM · AGENT

Competitor brief

drafted & appended to your log

5:11 AM · DIGEST

Inbox themes

☾ nothing new — stayed quiet

6:00 AM · DIGEST

Weekly themes across 41 new notes

ready

7:30 AM

4 updates waiting. 0 things you had to remember.

run now · pause & resume · a full run timeline for every automation

Four kinds, one habit

Pick your flavor of autopilot.

Underneath, every automation is the same promise — Chunk does the recurring thing so you don't have to. Four kinds cover almost anything you'd rather stop remembering.

RESEARCH

Standing research

Save a query once. Chunk re-runs it on schedule, diffs it against last time, and sends a cited digest of only what's new.

"Track solid-state battery breakthroughs" · weekly

Free: 1 · Pro: 5, daily runs, Deep Research

WATCHER

Page watchers

Point at any URL. Get pinged when the price dips below your number, it flips back in stock, or the page meaningfully changes.

"Tell me when this drops under $900" · checked daily

Free: 3 · Pro: 25

AGENT TASK⭐ Pro

Agent tasks

Write the job in plain English. A background agent reads your notes and the web, then writes the result into your knowledge base.

"Every Monday, brief me on my competitors"

Pro: 3 active

DIGEST

Note digests

No web at all — just your own notes, summarized on schedule: weekly themes, stale notes worth revisiting, connections you didn't spot.

"What did I actually capture this week?" · Sundays

Free: 1 weekly · Pro: 3

Same scheduling, same notifications, same filing — one substrate underneath.

Agent tasks · Pro

An agent that reads your notes before it reads the web.

Describe the job in one sentence. The agent pulls context from your notes, documents, collections, and past reports — then searches the web, does the work, and writes the result back where it belongs: a new note, an append to one you choose, or a report filed into a collection.

  • Plan preview see the numbered plan before anything runs.

  • Your context pick exactly which notes and collections it reads.

  • Working memory recurring runs skip what they already covered.

  • Any cadence once, daily, weekly, or monthly.

YOUR INSTRUCTION

“Every Monday morning, check what my competitors shipped and append a brief to my Competitor log.”

[[Competitors]] · collection12 notes2 past reports9.4k / 24k tokens

HERE'S WHAT I'LL DO

  1. 01Read 14 items from Competitors
  2. 02Search the web for launches & announcements this week
  3. 03Skip anything covered in last Monday's run
  4. 04Append a dated brief to "Competitor log"
Looks right — schedule itor edit

remembers what it covered between runs · Mondays · 7:00 AM

OR START FROM A RECIPE

Notes + web reportWeekly competitor briefTopic tracker logWeekly review writerReading triageTrip itinerary planner

The quiet part

Quiet until something changes.

Every run is compared against the last. Found something? A cited digest, instantly. Found nothing? Chunk says nothing at all — an automation that cries wolf gets deleted.

NEW 01

QuantumScape ships first commercial B-sample cells to a second OEM

A second automaker began validation testing this week — the clearest commercialization signal since the PowerCo deal.

reuters.com · electrek.co

NEW 02

Toyota pushes its solid-state timeline forward to 2027

Updated investor guidance moves mass production a full year earlier than the previous roadmap.

nikkei.com

NEW 03

New sulfide electrolyte paper claims 500+ cycle stability

Peer-reviewed results address the dendrite problem that has stalled the field for a decade.

nature.com · arstechnica.com

☾ No changes · Jun 24 — no email, no ping, no noise.
✉ Email on🔔 Push onset per automation

Nothing evaporates

A knowledge base, not a feed.

Automations don't dump into a timeline you have to babysit. Results land where your thinking already lives — searchable, linkable, citable, and there next month.

DIGEST

Your inbox

A cited email after any run that found something — toggleable per automation.

FILED

Your knowledge base

Reports file into collections. Agent tasks write into notes. Chat and search see every run.

PUSH

Your pocket

Push to iPhone, iPad & Mac the moment something real happens. Or silence — your call, per automation.

Email, push, both, or neither — chosen per automation, changed anytime.

Zero setup energy

Say it in one sentence.

Type what you want in plain English. Chunk works out which kind of automation it is, prefills the setup, and shows you exactly what it understood before anything runs.

Chat suggests automations mid-conversation — one click and it's running.

“Let me know when the Fujifilm X100VI is back under $1,600”

WATCHERprice + stock · checks daily
Price$1,749
In stockNo

read from the page just now

“…that's the third time you've asked about this launch. Want me to just watch it for you?”

Create automationsuggested in chat

Free to start

Free runs the errand. Pro runs the department.

Every kind has a free tier so you can feel the habit form. Pro turns it into staff.

FREE⭐ Pro
Standing research1 · weekly or monthly5 · daily · Deep Research depth
Page watchers325
Agent tasks3 active
Note digests1 · weekly themes3 · + stale notes & unseen connections

Push & email notifications on every plan. Quiet runs never ping — on any plan.

Stop remembering to check.

Set your first automation tonight — wake up to the answers.

Free: 1 research · 3 watchers · 1 digest — Pro adds daily runs, 25 watchers & agent tasks