decuff only talked to its own servers while we watched.E1
An AI agent used the site and saw 1 of 2 claimed features working.E37E40E41E39
AI / agents: I sent "hello" from the homepage box and "test" from the playground. Each time the page sent POST /v1/chat/completions, but no model answer came back, only "You're out of credits. Your message is kept." Getting a reply seems to need a sign-up or paid credits.E37E40E41
A working app / platform: Clicking ASK opened the chat playground, and its backend set up an account session (POST /v1/accounts, GET /v1/account) without asking for an email. I could not test the API for Claude Code and Codex because it needs a dc_sk_ API key. The docs list the endpoints and setup steps.E37E39
Also noticed, Status page loads live service data: The /status page called /api/uptime and the API catalog. It showed "All systems operational" with 2 models listed.E38
Tech check: the site claims ai / agents and a working app / platform, and its public page shows matching evidence.E32E33E13E14E15
We saw no known analytics or advertising services load on the first visit.E1
Tech check
What the site says about itself, next to what its public page actually loads. Facts only, no verdict: you decide.
Evidence found
AI / agents
“…Breach Docs Sign in Start a chat live · 2 models ready AI, decuffed. Uncensored open-weight models. The models n…”E32
“…you agree to the terms and confirm you’re 18 or older. API for Claude Code & Codex Models & pricing No email · no…”E33
Calls to its own backend: /v1/service, /v1/catalog, /assets/figures/cuff.jsonE13E14E15E16
Calls its own backend/v1/service, /v1/catalog, /assets/figures/cuff.jsonE13E14E15E16
Wallet codenone seen
Contract calls in its codenone seen
Talks to AI servicesnone seen
Live connectionsnone seen
What the AI agent tried
An AI agent (claude-opus-5-5) used the site in our sandbox for 7 steps and saw 1 of 2 claimed features working. It can't connect a wallet, pay, sign or log in, so those features count as "couldn't test". What it reports is what it observed, not a verdict.
Couldn't test
AI / agents
I sent "hello" from the homepage box and "test" from the playground. Each time the page sent POST /v1/chat/completions, but no model answer came back, only "You're out of credits. Your message is kept." Getting a reply seems to need a sign-up or paid credits.E37E40E41
Seen working
A working app / platform
Clicking ASK opened the chat playground, and its backend set up an account session (POST /v1/accounts, GET /v1/account) without asking for an email. I could not test the API for Claude Code and Codex because it needs a dc_sk_ API key. The docs list the endpoints and setup steps.E37E39
Also noticed
Things the agent saw along the way that aren't among the site's claims. Not counted above.
Noticed
Status page loads live service data
The /status page called /api/uptime and the API catalog. It showed "All systems operational" with 2 models listed.E38
Noticed
Free trial not applied to the anonymous playground session
The docs say accounts made without email start with a 500,000-token free trial on Flash. Even so, the first and second messages from the automatically created playground session both returned an out-of-credits notice.E37E39E41
What it is
The Internet Archive first saved this site on 2026-10-14 (28 days ago).E34
The site describes itself as: “Uncensored open-weight models that never refuse, on one OpenAI/Anthropic-compatible endpoint. Instant key, no KYC, private by design. Free t”E2
It's built with React (common tools for building web apps).E8
Notable building blocks in its code: Firebase SDK.E9
In your browser it can copy text to your clipboard and saves small bits of data in your browser.E10E11
It has sections such as /models, /breach, /docs, /account, /playground and /status.E31
While loading, it called 8 endpoints on its own servers, such as /v1/service, /v1/catalog, /assets/figures/cuff.json and 5 more.E13E14E15E16
Where your data goes
On a first visit the page only talked to its own servers (decuff.ai, api.decuff.ai).
How key features work
AI features
Parts of the site talk to AI services or AI endpoints (/v1/chat/completions). Text you type there may be sent to them.E26
/v1/chat/completions
Sign-in and accounts
Sign-in and account features show up in its code or traffic (/v1/account and /v1/account/sign-out).E29E30
/v1/account/v1/account/sign-out
Loading its content
Its content loads from its own API while you browse (/v1/service, /v1/catalog, /assets/figures/cuff.json, /assets/figures/slide.json and 2 more).E13E14E15E16E17
Every sentence above points to one of these. Code references show the file, line:column and the short identifier that matched, never the site's code itself.
E1
Page title: decuff · Uncensored open-weight AI — API for Claude Code & Codex
page title · https://decuff.ai/
E2
Page description: Uncensored open-weight models that never refuse, on one OpenAI/Anthropic-compatible endpoint. Instant key, no KYC, private by design. Free tokens to s
page description · https://decuff.ai/
E3
Site name: decuff
page site_name · https://decuff.ai/
E4
Main heading: AI, DECUFFED.
page h1 · https://decuff.ai/
E5
30 request(s) to decuff.ai
types: 15 script, 6 fetch, 5 font
GET https://decuff.ai/ · script · 30 requests · network recording
E6
2 request(s) to api.decuff.ai
types: 2 fetch
GET https://api.decuff.ai/v1/service · fetch · 2 requests · network recording
E7
localStorage: 1 key(s) saved in your browser
localStorage keys: decuff-theme · values not stored
E8
React detected by the code reader (react-runtime)
confidence: medium
decuff.ai/assets/index-BxwcazBZ.js · “react-runtime” · the code reader
The site says: “…Breach Docs Sign in Start a chat live · 2 models ready AI, decuffed. Uncensored open-weight models. The models n…”
page text · https://decuff.ai/ · …Breach Docs Sign in Start a chat live · 2 models ready AI, decuffed. Uncensored open-weight models. The models n…
E33
The site says: “…you agree to the terms and confirm you’re 18 or older. API for Claude Code & Codex Models & pricing No email · no…”
page text · https://decuff.ai/ · …you agree to the terms and confirm you’re 18 or older. API for Claude Code & Codex Models & pricing No email · no…
E34
The Internet Archive first saved decuff.ai on 2026-10-14 and has snapshots from 2 different days
page web archive · https://web.archive.org/web/20261014010133/decuff.ai
E35
Agent step 1: view
AI agent step 1 · view on /
E36
Agent step 2: type Ask anything, in full ← “hello”
AI agent step 2 · type on /
E37
Agent step 3: click ASK
AI agent step 3 · click on /playground · page called GET api.decuff.ai/v1/service, GET api.decuff.ai/v1/catalog, GET api.decuff.ai/v1/sandboxes/service, POST api.decuff.ai/v1/accounts, GET api.decuff.ai/v1/account, POST api.decuff.ai/v1/chat/completions
E38
Agent step 4: click STATUS
AI agent step 4 · click on /status · page called GET api.decuff.ai/v1/service, GET api.decuff.ai/v1/catalog, GET decuff.ai/api/uptime
E39
Agent step 5: go /docs
AI agent step 5 · go on /docs · page called GET api.decuff.ai/v1/service, GET api.decuff.ai/v1/catalog, GET api.decuff.ai/v1/account
E40
Agent step 6: go /playground
AI agent step 6 · go on /playground · page called GET api.decuff.ai/v1/service, GET api.decuff.ai/v1/catalog, GET api.decuff.ai/v1/account, GET api.decuff.ai/v1/sandboxes/service
E41
Agent step 7: type Message ← “test”
AI agent step 7 · type on /playground · page called POST api.decuff.ai/v1/chat/completions
Limits
We only see what your browser sees: code and requests on a first visit, without logging in or accepting cookie banners. What the servers do with data is invisible from outside.
Site code is usually minified (compressed), so names are partly lost and some code may be only partly analysed.
Some trackers only load after you accept cookies or log in, so the real list can be longer.
We visited from a server in the United States. Sites load different trackers and cookie banners in different countries (visitors in the EU usually get a consent banner first).
The tech check compares the site's own words with what its public page loads. Work done on servers, in private code or in separate apps isn't visible here, so 'nothing found' means not visible on the public page, not that it doesn't exist.
This text was written from fixed templates because no AI writer was available. It uses exactly the same evidence.
The code reader hit its size budget on parts of 5 large bundle file(s); those parts weren't fully analysed.
Facts, not judgements: this page describes what the code and traffic show, not anyone's intent. Not financial advice. Case id VYeThmfF.
Paste a coin address, a ticker, a wallet or an X post. The scan takes seconds, it's free, and there's no sign-up.
Every coin gets a green, yellow or red light with a score out of 100. Every point links to its source.
Is the tech real?90
Is the dev legit?40
Fair launch?20
2 of 4 · How to read a coin
Three questions, then the money
Tech, dev, launch: each has its own score and its top reasons, and the weakest one sets the light.
Who's behind it: the dev's linked wallets, crews that buy launches together, and who paid for the launch block.
Money taken out: what the dev and the launch-block wallets sold or pulled in the first 24 hours.
X@trenchinspector + CATelegram/s CADiscord/check CA
3 of 4 · Use it where you trade
Ask from X, Telegram or Discord
X: tag @trenchinspector with a coin address and it replies with the light and why.
Telegram and Discord: add the bot to your group and send a coin address.
Pulse: new launches appear about a second after they go live, and each one gets a light once it's scanned.
$COIN · dev sold 4.2%liquidity −40% in 10 min
4 of 4 · Stay on it
Track coins and wallets
Press Track on any coin or wallet. You'll get a browser notification when the dev sells, liquidity is pulled, the price moves hard, or a tracked wallet trades.
No account needed. Tracking is tied to this browser. See it all under Tracking.
