Serge

Serge — One-Click Install

LLaMA made easy

Why install Serge with Web.Casa?

One-Click Deploy

No manual Docker Compose needed. Web.Casa configures and starts everything.

Auto HTTPS

Caddy auto-provisions SSL certificates. Reverse proxy works out of the box.

Unified Management

Monitoring, backups, updates, logs — manage all apps from one panel.

How to Install

Install Web.Casa first, then one-click install this app from the App Store.

App Info

Version
0.9.0
Author
nsarrazin
Default Port
8008

Serge - LLaMA made easy 🦙

Serge is a chat interface crafted with llama.cpp for running Alpaca models. No API keys, entirely self-hosted!

  • 🌐 SvelteKit frontend
  • 💾 Redis for storing chat history & parameters
  • ⚙️ FastAPI + LangChain for the API, wrapping calls to llama.cpp using the python bindings

🧠 Supported Models

CategoryModels
Alpaca 🦙Alpaca-LoRA-65B, GPT4-Alpaca-LoRA-30B
Chronos 🌑Chronos-13B, Chronos-33B, Chronos-Hermes-13B
GPT4All 🌍GPT4All-13B
Koala 🐨Koala-7B, Koala-13B
LLaMA 🦙FinLLaMA-33B, LLaMA-Supercot-30B, LLaMA2 7B, LLaMA2 13B, LLaMA2 70B
Lazarus 💀Lazarus-30B
Nous 🧠Nous-Hermes-13B
OpenAssistant 🎙️OpenAssistant-30B
Orca 🐬Orca-Mini-v2-7B, Orca-Mini-v2-13B, OpenOrca-Preview1-13B
Samantha 👩Samantha-7B, Samantha-13B, Samantha-33B
Vicuna 🦙Stable-Vicuna-13B, Vicuna-CoT-7B, Vicuna-CoT-13B, Vicuna-v1.1-7B, Vicuna-v1.1-13B, VicUnlocked-30B, VicUnlocked-65B
Wizard 🧙Wizard-Mega-13B, WizardLM-Uncensored-7B, WizardLM-Uncensored-13B, WizardLM-Uncensored-30B, WizardCoder-Python-13B-V1.0

Additional weights can be added to the serge_weights volume using docker cp:

docker cp ./my_weight.bin serge:/usr/src/app/weights/

⚠️ Memory Usage

LLaMA will crash if you don't have enough available memory for the model:

ModelMax RAM Required
7B4.5GB
7B-q2_K5.37GB
7B-q3_K_L6.10GB
7B-q4_16.71GB
7B-q4_K_M6.58GB
7B-q5_17.56GB
7B-q5_K_M7.28GB
7B-q6_K8.03GB
7B-q8_09.66GB
13B12GB
13B-q2_K8.01GB
13B-q3_K_L9.43GB
13B-q4_110.64GB
13B-q4_K_M10.37GB
13B-q5_112.26GB
13B-q5_K_M11.73GB
13B-q6_K13.18GB
13B-q8_016.33GB
33B20GB
33B-q2_K16.21GB
33B-q3_K_L19.78GB
33B-q4_122.83GB
33B-q4_K_M22.12GB
33B-q5_126.90GB
33B-q5_K_M25.55GB
33B-q6_K29.19GB
33B-q8_037.06GB
65B50GB
65B-q2_K29.95GB
65B-q3_K_L37.15GB
65B-q4_143.31GB
65B-q4_K_M41.85GB
65B-q5_151.47GB
65B-q5_K_M48.74GB
65B-q6_K56.06GB
65B-q8_071.87GB

🧾 License

Nathan Sarrazin and Contributors. Serge is free and open-source software licensed under the MIT License.

💬 Support

Need help? Join our Discord

FAQ

Can I install Serge on Web.Casa?

Yes, Serge is available in the Web.Casa App Store. You can install it with one click from the Web.Casa panel — no need to manually write Docker Compose files or configure reverse proxies.

How do I install Serge on Web.Casa?

First install Web.Casa with curl -fsSL https://get.web.casa | bash, then search for "Serge" in the App Store and click install. Web.Casa automatically configures the Docker container, assigns ports, and sets up a Caddy reverse proxy.

What is the default port for Serge on Web.Casa?

The default port for Serge on Web.Casa is 8008. You can bind it to a custom domain through Web.Casa's Caddy reverse proxy, which automatically provisions HTTPS certificates.

Does Serge get automatic HTTPS after installation?

Yes. Web.Casa is built on Caddy, so after installing the app, simply bind a domain and Caddy will automatically provision and renew SSL certificates via Let's Encrypt — no manual configuration needed.

How do I back up Serge data?

Web.Casa has a built-in backup plugin that supports automatic backups of Serge data to local storage, S3, WebDAV, or SFTP. You can configure scheduled snapshots and retention policies to keep your data safe.