English · Español · Français · Deutsch · Português · Italiano · Türkçe · 한국어 · 日本語 · 简体中文 · Русский · Polski · हिन्दी · Nederlands · Українська · Tiếng Việt · Bahasa Indonesia · ไทย · فارسی · Ελληνικά · עברית · Čeština · Română · Svenska · Dansk · Suomi · Magyar · العربية · Oʻzbekcha · Azərbaycan · Български · Hrvatski · Қазақша · ქართული · नेपाली · සිංහල · தமிழ் · తెలుగు · मराठी · বাংলা · ខ្មែរ · Lietuvių · Bahasa Melayu · မြန်မာ · Norsk · Slovenčina · Slovenščina · Српски · Kiswahili · Filipino · اردو
Uncle Sam
A DeepSeek Harness provider that answers through your own browser, on your existing web account, instead of the billed API.
The DeepSeek Harness plugin that drives chat.deepseek.com in a real browser — a free DeepSeek API alternative. Chat, DeepThink and web search on your own account: no API key, no cost per token, no billing.
☕ Buy me a coffee — written by one person, at night, funded by nothing. There is no other revenue model. A star or a bug report is worth just as much.
This product is ongoing and under active development. Expect bugs.
It works, it is fast, and it has been driven daily for weeks — but it keeps changing, and the browser route
depends on a website that changes without telling anyone. If something breaks, a screenshot and the version
number is worth more than a description. Pin a version if you depend on it.
provider deepseek-web
models deepseek-chat · deepseek-chat-search · deepseek-reasoner · deepseek-reasoner-search
cost nothing per token — it uses the subscription already open in your browser
status ongoing development — expect bugs
measured first token ~1.9 s · overhead ~0.6 s · a full turn in ~2.5 s
Replies from the web route are signed 🔴 Uncle Sam. That name is a joke, not a claim: it marks which route
answered, and API replies carry no signature at all. You always know which of the two answered you.
What it saves you
The API bills per token. The web product does not. This plugin registers deepseek-web as a model provider and
answers every call by driving chat.deepseek.com in a real browser — the page does the work, the browser holds
the session, and nothing is metered.
| API route | Web route |
|---|
| Cost | per token, forever | nothing beyond the subscription |
| Speed | fast | ~2.5 s a turn, measured |
| Tool calling | native and reliable | emulated, works but weaker |
| Depends on | the API | a website that changes without notice |
Both routes stay installed. The model picker is the switch, so you can use the cheap one by default and
reach for the API when a task needs reliable tool use.
Installing
Uncle Sam is a DeepSeek Harness plugin. It does nothing outside the harness, because it registers a model
provider, a host route and two client UI slots.
Requirements
| |
|---|
| DeepSeek Harness | any recent build with the plugin manager |
| Browser | Brave recommended, Edge used automatically when Brave is absent |
| Node | 22 or newer |
From npm
pnpm add @operatingtag/dsh-unclesam
From this repository
git clone https://github.com/operatingtag/dsh-unclesam
cd dsh-unclesam
pnpm install
Then add the bundle row to your profile so the harness loads it:
# cordis.patch.yml
- insert:
- id: dsh-unclesam
name: '@operatingtag/dsh-unclesam'
config:
port: 9222
defaultModel: deepseek-chat-search
autoMinimise: true
Or from a folder: Settings → Plugins → install from a local path, pointed at the folder holding
package.json.
First run
Click Sign in / open chat. The browser opens and stays visible until you have signed in at
chat.deepseek.com. After that it minimises itself and you will not see it again. The session lives in your
browser profile, so no password is stored by the plugin — there is nothing to leak.
You should then see:
- Settings → Uncle Sam — a page with status rows and a sign-in button.
- Under the chat box —
● DeepSeek Harness or ● Uncle Sam, click to switch.
- The model list — four
(web) models beside the API models.
Removing it
Settings → Plugins → Uncle Sam → Uninstall. Nothing is written outside the browser profile you signed in to.
The API / web switch
The model picker is the switch. There is no second control, because a second control that also chose the
model would be two sources of truth for one decision.
| Pick this model | Route | DeepThink | Search |
|---|
DeepSeek Chat (web) | browser | off | off |
DeepSeek Chat + Search (web) | browser | off | on |
DeepSeek Reasoner (web) | browser | on | off |
DeepSeek Reasoner + Search (web) | browser | on | on |
any model without (web) | the API | — | — |
A WEB or API badge renders beside the model label so the active route is visible at a glance. When the
label is not recognised the badge renders nothing rather than guessing — a wrong badge is worse than none.
Your choice persists across restarts. DeepThink is the reasoning pass; Search lets it read the live web.
Search is visible in the transcript
With a search model, a turn produces two blocks — the query it searched with, the sources it found, then
the answer that used them:
Web search — "What is one technology news headline from today?"
Sources (4): yna.co.kr · fnnews.com · rthk.hk · sina.com.cn
Links: https://k.sina.com.cn/article_... http://www.21jingji.com/article_...
The page does not expose the query it typed internally, so the quoted query is the prompt that triggered the
search. Sources and links are read from the page once the answer settles. Because sources only exist at the
end, search turns buffer instead of streaming; ordinary turns still stream token by token.
Why a browser, and not their HTTP API
The obvious route — calling chat.deepseek.com's own API — was attempted and abandoned. The site makes every
message clear two gates:
| Gate | What it means |
|---|
/api/v0/chat/create_pow_challenge | a proof-of-work challenge per message |
challenges.cloudflare.com/turnstile | a bot gate on the origin |
Re-implementing either is a losing game: they are obfuscated, they change, and when they change the failure
looks exactly like a network fault. The page already solves both for its own requests.
So the adapter never speaks HTTP to the site. It drives the page, and the page does the rest. Auth,
proof-of-work and Turnstile stop being our problem and cannot break independently of the site itself.
Verified: a plain HEAD to the site returns 403; the browser path returns answers.
Limits, stated honestly
Tool use is emulated. A web chat has no structured tool channel, so tools are described in the prompt and
a call is recovered by parsing a JSON object out of the reply. It works — verified — but it is weaker than
native tool calling: a model that ignores the instruction answers in prose, and the harness sees no tool use
that turn. For heavy agentic work the API route is more reliable. The web route is best for chat, long
answers, subagents, compaction, and any turn where the model must be the web model.
It depends on a website you do not control. When the page changes shape, the plugin has to change with it.
That is the ongoing-development warning at the top of this page, not a formality.
It is not affiliated with DeepSeek. It is not endorsed by them, and it gives you no rights in their
service. Your use of that service is governed by their terms.
Configuration
cordis.patch.yml sets:
| key | default | meaning |
|---|
port | 9222 | the browser's debugging port |
defaultModel | deepseek-chat-search | used when a call names no model |
showBrowser | true | whether the window is visible while it works |
The browser must hold a signed-in chat.deepseek.com session. The first call launches Brave on the Default
profile if the port is not already open. If Brave is already running without a debugging port, the adapter
cannot attach, and says so clearly rather than failing silently.
Architecture
index.js the adapter: harness StreamChunk ←→ engine
lib/engine.mjs browser + page: type, send, stream, settle
lib/cdp.mjs Chrome DevTools Protocol client, zero dependencies
lib/prompt.mjs flatten messages + tools into one message; recover tool calls from the reply
lib/savings.mjs what the web route saved you, priced against the API
lib/i18n.mjs 51 languages for the panel
No dependencies. Node 24 ships a global WebSocket and fetch, which is all CDP needs, so the token
stream never leaves the host process and there is no IPC hop per delta.
Support
Uncle Sam exists so you stop paying per token. If it saved you money, and you would like to see the next one
built on hardware that does not run out of memory halfway through:
☕ revolut.me/operatingtag
A star on the repo or a useful bug report is worth just as much.
Licence
CC BY-NC-ND 4.0 — Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International.
Copyright (c) 2026 Darren T. Weldon, published under the Operating Tag name.
The whole deal, in one line: install it and use it, personally, for free. Do not change it, do not share
it, do not sell it, and do not use it in a business.
That is not open source, and it does not pretend to be.
you may install it and run it · use it personally, non-commercially
read the source to understand it
you may not modify, adapt or create a derivative of it
copy, distribute, publish, mirror or make it available to anyone
upload it to any code host, registry, website, forum or chat
sell, resell, rent, or charge for it or for access to it
bundle it into anything you supply to another person
use it for any commercial purpose, or in the course of a business
The only two legitimate sources are this repository and the npm package
@operatingtag/dsh-unclesam. A copy from
anywhere else is not licensed, whoever supplied it. If you see this software for sale, it is not being sold by
its author.
Commercial use, and the right to modify and distribute, are available under a separate written licence.
Ask Operating Tag and you will get a straight answer and a clear price.
The LICENSE file governs, and it opens with the same summary in plain English.
Uncle Sam is an independent plugin. Not affiliated with, endorsed by, or supported by DeepSeek.
@operatingtag/dsh-unclesam v1.3.5
Author Operating Tag
Licence CC BY-NC-ND 4.0 — free to use, not to change, share or sell
