ads
Almost no paid ads
One LinkedIn ad. No Meta ads for its name.
LinkedIn and Meta ad libraries via ScrapeCreators · 30 Oct
looot /gtm
Growing your GTM only takesone command.
One domain in. Accounts, buyers, rivals and ranks out. You see the price before anything paid runs.
/gtm yourdomain.com
Claude Code
Codex
Cursor- any terminal
/gtm opens soon. It runs on your looot key.
answered in the TypeSafe run
Jina Reader
Hunter
BuiltWith- AnyAPI
- Datagma
DataForSEO
ScrapeCreators
The Companies API
LeadMagic
Anysite- Cloro
Out comes your whole GTM.
Four steps. The first is free. Nothing paid starts before your yes.
brief
typed decisions for software
accounts
Vercel, Cloudflare, LangChain
people
nobody on file yet
$ /gtm yourdomain.comrival cards
rival launches in Oct
10-0110-1210-22
ads
one LinkedIn ad
Search + AI
#1 for jev ai model
typesafe.ai
Reads 4 pages. Free.
gtm/config.yamlPrints the price.
Waits for your yes.
Runs inside one limit.
gtm/dossier.html
The four steps in full
- It reads your site for free. It reads 4 pages of your site with a plain web request. No looot call and no charge. Then it writes a short brief: what you sell, who buys it, the customers you want more of, your rivals, your keywords and the questions buyers ask. Change anything you like.
- It shows you the price. It prints every module with its calls, the expected cost, the most it could cost and a suggested spending limit. Nothing paid has run yet.
- It waits for your yes. It asks one question. Run it with this limit? Only a yes starts the paid run. Any other answer stops it and nothing is spent.
- It runs inside one limit. It runs every module under that one limit. Then it writes the dossier and a receipt of every paid call.
Fifteen parts. One report.
Logo: who answered. Dashed: skipped, free.
Intake
Jina ReaderIdeal customer
HunterAccounts
The Companies APIPeople
LeadMagicWork emails
skipped, $0Signals
AnyAPICompetitors
AnyAPIAds
DataForSEOSearch + AI
AnyAPIReviews
skipped, $0Mentions
ScrapeCreatorsMarket
DatagmaLocal
skipped, $0Commerce
skipped, $0Summary
What each part hands back
- Intake.
- Reads your home, pricing, customers and comparison pages and drafts your brief.
- Ideal customer.
- Builds your ideal customer profile and finds lookalike companies from your seed customers.
- Accounts.
- Finds companies that fit and grades each one from A to D with its reasons.
- People.
- Finds the buyers at each account by title.
- Work emails.
- Finds their work emails and checks each one with a verifier. Phones only when you ask.
- Signals.
- Hiring, funding and news per account, each with a date and a hot, warm or cold score.
- Competitors.
- One card per competitor. Every claim on it has a source and a date.
- Ads.
- The ads you and your competitors run, with their angles and offers.
- Search + AI.
- Keyword gaps, your Google ranks and how often AI answers name you.
- Reviews.
- Review themes and competitor weak spots from the App Store and Google Maps.
- Mentions.
- Posts from the last week and the ones that show buying intent.
- Market.
- Traffic, funding, hiring and news for you and each competitor.
- Local.
- Runs when people visit your business. Your rank and rating in local search.
- Commerce.
- Runs when you sell on Amazon. Price, rating and reviews per listing.
- Summary.
- Names your hot accounts, the rival ad angles you do not use yet and where each rival is weak.
We ran it on TypeSafe AI.
A live run on 30 Oct 2026.
typesafe.ai
TypeSafe AI
Jev returns typed decisions that software can act on.
Out of stealth on 15 Sep 2026. Series A less than four weeks later.
- staff on file
- 11-50
Hunter company profile · 30 Oct
Here is how it grew.
Total raised and every dated event the run found.
total raised
$910M
24 days from seed to Series A13 dated events
swipe the chart
Sources the run returned
All sources, with dates
15 Sep · funding Out of stealth with a $40M seed
16 Sep · press Debuts a model for machines
18 Sep · press A new kind of AI model is thrilling developers
19 Sep · press Vercel and Cloudflare rushed to add it
19 Sep · community Reddit post "Jev is revolutionary", 194 upvotes
22 Sep · press Why everyone is talking about Jev
1 Oct · rivals A large cloud ships a free open source rival
9 Oct · funding Series A at a $7.5B valuation
12 Oct · rivals Decision models flood the market
14 Oct · hiring Hiring a Model Capabilities Engineer
15 Oct · hiring Hiring a Software Engineer Generalist
22 Oct · rivals A large lab makes its decision model widely available
27 Oct · community An open source Jev clone, 216 upvotes
What the run found on TypeSafe.
One finding per part of the report, each with its source.
seo
An AI answer missed it
A buyer question got a rival lab's API instead of Jev.
ChatGPT answer via Cloro · 30 Oct
competitors
Rivals arrived within weeks
Large labs and clouds shipped rival decision models in October.
signals
Hiring engineers right now
Two engineering roles in San Francisco, posted mid October.
LinkedIn jobs via AnyAPI · 15 Oct
icp
The stack on file
Next.js, React, TypeScript, Python, Kubernetes, Kafka, Terraform and AWS.
Hunter company profile · 30 Oct
people
No people on file yet
One people call found nobody yet. The empty answer cost $0.
LeadMagic, one call · 30 Oct
Who builds with it
- Vercel200-500 staff
- Cloudflare1k-5k staff
- LangChain10-50 staff
- Jack & Jill51-250 staff
The Companies API · 30 Oct
Search + AI
jev ai model
- 1typesafe.ai
- 2en.wikipedia.org
- 3typesafe.ai
- 4greennode.ai
- 5langchain.com
Google results via AnyAPI · 30 Oct
AI answer
Which model for fast typed decisions in an agent?
Jev not named. A rival lab's API was.
ChatGPT answer via Cloro · 30 Oct
What the run could not show
- The traffic provider's newest snapshot is from May 2026, before the launch, so this page does not chart traffic.
- Two lookalike searches came back thin. The accounts above are the ones the sources name.
TypeSafe AI has no link to looot. This is public data, read by looot on 30 Oct 2026.
You see the price first. You set a limit.
What the TypeSafe run cost, against its quote and its limit.
charged · typesafe.ai
$0.11
spending limit
$3.00
expected $0.13most it could cost $0.42
$0$3.00
zoomed to the most it could cost
charged $0.11expected $0.13
$0most it could cost $0.42
How the spending limit works
- quote
- share per part
- hold before each call
- stop at the limit
One report. You see every call and what it cost.
Every call, its provider and its charge. Misses stay in.
typesafe.ai · 2026-10-30
- web.scrape.markdownJina Readerfound$0.0001
- company.enrichHuntersecond try$0.0049
- company.technographicsBuiltWithfound$0
- news.searchAnyAPIfound$0.0010
- google.serp.organicAnyAPIfound$0.0005
- company.trafficDatagmafound$0.065
- keywords.domain.overviewDataForSEOfailed$0
- keywords.domain.overviewDataForSEOfound$0.012
+17 more calls
| part | job | provider | outcome | charged | run id |
|---|---|---|---|---|---|
| intake | web.scrape.markdown | Jina Reader | foundtypesafe.ai home page | $0.0001 | run_a06f46e869584160ae744f4bef06387f |
| icp | company.enrich | Hunterafter The Companies API | found on the second trytypesafe.ai | $0.0049 | run_211f6877e89c48ebadcd60ee11089041 |
| icp | company.technographics | BuiltWith | foundtypesafe.ai | $0 | run_d90dedfd44b346758a3a8204eb0a7f7f |
| signals | news.search | AnyAPI | foundTypeSafe AI Jev | $0.0010 | run_577f0017610b460281a64bc37fe3be3f |
| seo | google.serp.organic | AnyAPI | foundjev ai model | $0.0005 | run_d2c9df79b7cb4410adce557f8ed3a92b |
| market | company.traffic | Datagma | foundtypesafe.ai | $0.065 | run_865a462cfd6e4de5b50e77f2204bbaa8 |
| seo | keywords.domain.overview | DataForSEO | failedrejected before the call: wrong input field, no charge | $0 | run_6352834cfbfc4719b79876ab1f9ed177 |
| seo | keywords.domain.overview | DataForSEO | foundtypesafe.ai | $0.012 | run_591b901fde8c4472bf8ce5e7fbf3d61b |
| mentions | reddit.post.search | ScrapeCreators | foundJev TypeSafe | $0.0019 | run_9a1ec9f4ca1945eb94a4d878ce2c0457 |
| ads | ads.google.search | DataForSEO | failedrejected before the call: wrong input field, no charge | $0 | run_c87499ee830248db85e085e06a22974f |
| ads | ads.meta.search | ScrapeCreators | foundTypeSafe AI: no ad from TypeSafe in the results | $0.0019 | run_d5ff278b9a454ff0974c804c753d235d |
| ads | ads.linkedin.search | ScrapeCreators | foundTypeSafe AI | $0.0019 | run_a65e7b72d27c42808c6bf6e7904c5c88 |
| accounts | company.similar | The Companies API | came back emptytypesafe.ai: empty list, not billed | $0 | run_61b27fcdb58f465faba82b1e901ca4bf |
| competitors | news.search | AnyAPI | founddecision model rivals | $0.0010 | run_657fa714955a4af4b0c52f377d8c815b |
| signals | jobs.search | AnyAPI | failedrejected before the call: wrong input field, no charge | $0 | run_dddb9aeb65ff417988b7ccda8a5601a5 |
| people | people.search | LeadMagic | came back emptytypesafe.ai: nobody on file yet; a first provider also came back empty; not billed | $0 | run_1bdaf97b63f34d3bb7bbf064b0a223b3 |
| signals | jobs.search | AnyAPI | foundTypeSafe AI | $0.0006 | run_a819634db77d4242a84613fdb2fc89f3 |
| accounts | company.search | Anysite | came back emptyempty answer, billed | $0.0036 | run_7f2f788ef93d413eb3d57a8d742c7988 |
| seo | chatgpt.search.answer | Cloro | failedrejected before the call: missing country, no charge | $0 | run_650ebcbd613a40b286d078a0397d591d |
| seo | chatgpt.search.answer | Cloro | foundbuyer question about typed decisions | $0.0056 | run_a84605e6be564e62be8a9be7486022ea |
| accounts | company.search | Anysite | foundAI agent platform | $0.0036 | run_d0ea6285d72a432fa778a68354c20317 |
| accounts | company.enrich | The Companies API | foundvercel.com | $0.0019 | run_5d11e7c31c1b4680b4fbb7f83380a908 |
| accounts | company.enrich | The Companies API | foundcloudflare.com | $0.0019 | run_6adb3eb4e8954003a42db7ce68ddcc1e |
| accounts | company.enrich | The Companies API | foundlangchain.com | $0.0019 | run_db435f21ef1543ad9a6b4dcd700896b3 |
| accounts | company.enrich | Hunterafter The Companies API | found on the second tryjackandjill.ai | $0.0049 | run_99158ad2c1e4448082470b1e3a9ca186 |
receipt
$0.11
- 25 calls, 18 billed, the rest cost nothing
- done in 3 min 48 s
- spending limit $3.00
- Intake
- Ideal customer
- Accounts
- People
- Work emails
- Signals
- Competitors
- Ads
- Search + AI
- Reviews
- Mentions
- Market
- Local
- Commerce
- Summary
Many providers. One key, one receipt.
Without /gtm
11 keys, 11 bills
With /gtm
one looot keyone receiptLists ready for the tools you use.
Import files with verified emails only.
Instantly
exports/instantly.csvVerified emails only
HubSpot
exports/hubspot.csvVerified emails only
Attio
exports/attio.csvVerified emails only
What lands in your gtm folder
- dossier.html
- dossier.md
- accounts.csv
- people.csv
- receipt.json
- changes.md
Pop it into Claude Code, Codex or Cursor.
The /gtm skill shows the price, waits for your yes, then runs.
# The skill file goes in
~/.claude/skills/gtm/SKILL.md
> /gtm yourdomain.comClaude Code reads skills here. Start a session, type /gtm and your domain.
What your agent does
- Free read and quote
- Shows you the price
- Asks and waits
- Runs inside your limit
- Hands you the dossier
Rules your agent follows with your money.
Written into the skill file your agent reads. Tap one for the detail.
Price first
Every command that spends money shows the price first and asks. The quote and the demo never spend.
No yes, no run
Your agent never skips the question before you said yes to that limit. Without a terminal to answer in, the command stops after the quote and spends nothing.
The limit is yours
Your agent never raises the spending limit on its own. If the run needs more, it asks you for a higher limit or a smaller run.
Two failures, then stop
If a command fails twice, the agent stops and shows you the error. It does not try other flags, other domains or other tools.
Your key stays in your shell
The agent never asks you to paste your looot key into the chat. You set it in your shell before you start.
Questions before you try it.
One domain. It reads 4 pages of your site and drafts the brief. You can fix the brief before anything paid runs.
No. It writes research and copy. You check each one, then you send the emails and run the ads yourself.
It prints the quote before any paid call: calls per module, the expected cost, the worst case and a suggested limit. Prices come from the looot catalog at that moment. The TypeSafe run on this page is one real example.
It holds the most each call can cost before making it. It skips any call that could take the run past your limit. When it reaches the limit, it stops and keeps what it found.
A miss is a normal answer and stays in the dossier as a row. Some providers bill a not found. The receipt shows the real charge on every row.
It writes CSV files ready for Instantly, HubSpot and Attio, with verified emails only. A plain CSV works with anything else.
Yes. Demo mode writes a full dossier from made up data. No key, no network and no charge.
Yes. The public export swaps your domain, your rivals and every person for stand ins. Then it searches the files for every original name. If one is left, it deletes the copy and tells you which.
Claude Code, Codex and Cursor through the /gtm skill, or any terminal on its own.
Yes. Each run compares itself with the last one and writes what changed: new accounts, rival changes, rank moves, new signals and ad changes.
Give your agent one key.
Start your looot workspace. The same key pays for every provider in a /gtm run. You see each price first.