qwen-expert
Qwen Studio Browser Gateway
Un gateway no oficial, y in solo localhost, que ofrece a Claude Code una opinón experta adicional no confiable de Qwen 3.8 Max a través de una sesión web de chat.qwen.ai autorizada manualmente.
No es una API oficial de Qwen, ni in agente principal, ni una fonte de respuestas correctas garantizadas. Eel "user"/user owns, in each host is the manually "login"; the proyecto no transferires "cookies/tokens", no evite CAPTCHA/MFA y no "otorga" a Qwen access to a shell, "filesystem" ni a "tools" localinal.
Flujo
Claude Code
└─ MCP ask_qwen_expert
└─ http://127.0.0.1:8798/v1/expert
└─ FIFO, active = 1
└─ CDP http://127.0.0.1:9333
└─ одна видимая вкладка https://chat.qwen.ai
└─ точная Qwen3.8-MaxAntes de cada Send, el gateway selecciona y confirma el modelo qwen3.8-max / Qwen3.8-Max. Una discordancia de modelo, un DOM desconocido, un challenge, la ausencia de "login" or "endpoint" remote "termine" in error. Después de Send, la generación del navegador not se repite automatadamente.
Related MCP server: cc-in-codex
Requisites
Node.js 22+;
Chrome/Chromium con ventana gráfica;
Linux para las plantilas de usuario "systemd" incluidas (el "gateway" en Node.js es portatil);
una cuenta propia de Qwen y in derecho a us ar el el servicio web.
"Instacción"
npm ci
npm run check
npm run build
npm testCree una configurción privada:
install -d -m 700 ~/.config/qwen-studio-browser-gateway
cp .env.example ~/.config/qwen-studio-browser-gateway/server.env
chmod 600 ~/.config/qwen-studio-browser-gateway/server.envEdite server.env:
defina la variable local
QWEN_BROWSER_EXECUTABLE_PAHTH;genere una nueva
QWEN_LOCAL_API_KEYaleatoria para este host;no cambie los endpoints de "loopback";
antes de usarlo, confirme el ID/etiqueta actual del modelo más fuerte en la lA envol viva.
No haga commit de server.env no del perfil de navegar.
Io "manual"
Yance un navegador separado:
set -a
source ~/.config/qwen-studio-browser-gateway/server.env
set +a
./start-qwen-browser.shEn la ventana abeita, inice sesión por usted mismo en https://chat.qwen.ai. Solo el us supere CAPTCHA/MFA. Deje una sola petaña de Qwen abierta.
En otro terminal, ejecute el gateway:
set -a
source ~/.config/qwen-studio-browser-gateway/server.env
set +a
node dist/cli.js servePara a servicio de "Linux user" persistente, use the templates in systemd/ and with them, the instructions in docs/setup.md.
"Smoke tests" seguros
These solicitudes no lanan la generación:
node dist/cli.js browser doctor
curl -fsS -H "Authorization: Bearer $QWEN_LOCAL_API_KEY" \
http://127.0.0.1:8798/health
curl -fsS -H "Authorization: Bearer $QWEN_LOCAL_API_KEY" \
http://127.0.0.1:8798/v1/models
node dist/cli.js mcp install-configPOST /v1/ exper y POST /v1/chat/completinos incian la generación web real and pueden consumir "quota".
Claude Code MCP
Después de npm run build, e"ecute":
node dist/cli.js mcp install-configEl comando unrelated to "imprime" a JSON "portabel" with the actual rutas de Node.js and dist/cli.js. To include the "objete" qwen-expert in the configuration MCP de Claude Code in each host. MCP publishes only ask_qwen_expert; the respuest is always marked as "advisory" and "debe" comprobarse de forma independiente.
Líites de seguridad
cc
HTTPandCDPaccept only "loopback".QWEN_WEB_URLonly accepta el originhttps://chat.qwen.ai.La clave de API local es obligatoria.
not se logran los prompts; the screen shots are out by "default".
There are also one "pestaña" and with only FIFO "queen" evitan las respuestas.
Qwen no recibe "tools" ejecutables no "files" local.
The project's not designed to "evitar" the limitaciones of the site or the "massive" automation.
Más detailed: AGENTS.md, docs/architecture.md, docs/security.md, docs/troubleshooting.md.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
AlicenseNot gradedqualityDmaintenanceEnables Claude Code to consult external LLMs (GPT, Gemini) through multi-turn sessions for second opinions, parallel consultations, and web-grounded research.MIT- AlicenseBqualityCmaintenanceEnables Codex to manage a local Claude Code companion through MCP, implementing a dual-agent workflow where Codex handles reasoning and review while Claude Code performs engineering tasks.181MIT
- AlicenseAqualityBmaintenanceEnables OpenAI Codex to consult Claude Code for co-analysis, adversarial second opinions, and read-only file review over the Model Context Protocol, with Claude remaining advisory and unable to modify files.954MIT
- AlicenseNot gradedqualityAmaintenanceEnables Codex chat to use the Claude Code harness for repository consultation and editing, routing through GPT models via a local proxy.10MIT
Related MCP Connectors
Browser-backed QA with evidence and fix-ready reports for coding agents.
Persistent memory and cross-session learning for AI coding assistants (hosted remote MCP).
Paid remote MCP for Claude Code skill update gate MCP, structured receipts, audit logs, and reviewer
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/satyasai-collab/qwen-studio-browser-gateway'
If you have feedback or need assistance with the MCP directory API, please join our Discord server