containers/ramalama: The goal of RamaLama is to make working with AI boring.

The RamaLama project’s goal is to make working with AI boring
through the use of OCI containers.

RamaLama tool facilitates local management and serving of AI Models.

On first run RamaLama inspects your system for GPU support, falling back to CPU support if no GPUs are present.

RamaLama uses container engines like Podman or Docker to pull the appropriate OCI image with all of the software necessary to run an AI Model for your systems setup.

Running in containers eliminates the need for users to configure the host system for AI. After the initialization, RamaLama runs the AI Models within a container based on the OCI image.

RamaLama then pulls AI Models from model registries. Starting a chatbot or a rest API service from a simple single command. Models are treated similarly to how Podman and Docker treat container images.

When both Podman and Docker are installed, RamaLama defaults to Podman, The RAMALAMA_CONTAINER_ENGINE=docker environment variable can override this behaviour. When neither are installed RamaLama will attempt to run the model with software on the local system.

RamaLama supports multiple AI model registries types called transports.
Supported transports:

RamaLama uses the Ollama registry transport by default. Use the RAMALAMA_TRANSPORTS environment variable to modify the default. export RAMALAMA_TRANSPORT=huggingface Changes RamaLama to use huggingface transport.

Individual model transports can be modifies when specifying a model via the huggingface://, oci://, or ollama:// prefix.

ramalama pull huggingface://afrideva/Tiny-Vicuna-1B-GGUF/tiny-vicuna-1b.q2_k.gguf

To make it easier for users, RamaLama uses shortname files, which container
alias names for fully specified AI Models allowing users to specify the shorter
names when referring to models. RamaLama reads shortnames.conf files if they
exist . These files contain a list of name value pairs for specification of
the model. The following table specifies the order which RamaLama reads the files
. Any duplicate names that exist override previously defined shortnames.

$ cat /usr/share/ramalama/shortnames.conf
[shortnames]
  "tiny" = "ollama://tinyllama"
  "granite" = "huggingface://instructlab/granite-7b-lab-GGUF/granite-7b-lab-Q4_K_M.gguf"
  "granite:7b" = "huggingface://instructlab/granite-7b-lab-GGUF/granite-7b-lab-Q4_K_M.gguf"
  "ibm/granite" = "huggingface://instructlab/granite-7b-lab-GGUF/granite-7b-lab-Q4_K_M.gguf"
  "merlinite" = "huggingface://instructlab/merlinite-7b-lab-GGUF/merlinite-7b-lab-Q4_K_M.gguf"
  "merlinite:7b" = "huggingface://instructlab/merlinite-7b-lab-GGUF/merlinite-7b-lab-Q4_K_M.gguf"
...

RamaLama is available via PyPi https://pypi.org/project/ramalama

Install RamaLama by running this one-liner:

curl -fsSL https://raw.githubusercontent.com/containers/ramalama/s/install.sh | bash

You can run a chatbot on a model using the run command. By default, it pulls from the Ollama registry.

Note: RamaLama will inspect your machine for native GPU support and then will
use a container engine like Podman to pull an OCI container image with the
appropriate code and libraries to run the AI Model. This can take a long time to setup, but only on the first run.

$ ramalama run instructlab/merlinite-7b-lab
Copying blob 5448ec8c0696 [--------------------------------------] 0.0b / 63.6MiB (skipped: 0.0b = 0.00%)
Copying blob cbd7e392a514 [--------------------------------------] 0.0b / 65.3MiB (skipped: 0.0b = 0.00%)
Copying blob 5d6c72bcd967 done  208.5MiB / 208.5MiB (skipped: 0.0b = 0.00%)
Copying blob 9ccfa45da380 [--------------------------------------] 0.0b / 7.6MiB (skipped: 0.0b = 0.00%)
Copying blob 4472627772b1 [--------------------------------------] 0.0b / 120.0b (skipped: 0.0b = 0.00%)
>

After the initial container image has been downloaded, you can interact with
different models, using the container image.

$ ramalama run granite3-moe
> Write a hello world application in python

print("Hello World")

In a different terminal window see the running podman container.

$ podman ps
CONTAINER ID  IMAGE                             COMMAND               CREATED        STATUS        PORTS       NAMES
91df4a39a360  quay.io/ramalama/ramalama:latest  /home/dwalsh/rama...  4 minutes ago  Up 4 minutes              gifted_volhard

You can list all models pulled into local storage.

$ ramalama list
NAME                                                                MODIFIED     SIZE
ollama://smollm:135m                                                16 hours ago 5.5M
huggingface://afrideva/Tiny-Vicuna-1B-GGUF/tiny-vicuna-1b.q2_k.gguf 14 hours ago 460M
ollama://moondream:latest                                           6 days ago   791M
ollama://phi4:latest                                                6 days ago   8.43 GB
ollama://tinyllama:latest                                           1 week ago   608.16 MB
ollama://granite3-moe:3b                                            1 week ago   1.92 GB
ollama://granite3-moe:latest                                        3 months ago 1.92 GB
ollama://llama3.1:8b                                                2 months ago 4.34 GB
ollama://llama3.1:latest                                            2 months ago 4.34 GB

You can pull a model using the pull command. By default, it pulls from the Ollama registry.

$ ramalama pull granite3-moe
 31% |████████                    |  250.11 MB/ 783.77 MB  36.95 MB/s       14s

You can serve multiple models using the serve command. By default, it pulls from the Ollama registry.

$ ramalama serve --name mylama llama3

You can stop a running model if it is running in a container.

To use a UI, run a ramalama serve command, then connect via your browser at:

127.0.0.1:8080

+---------------------------+
|                           |
| ramalama run granite3-moe |
|                           |
+-------+-------------------+
	|
	|
        |           +------------------+           +------------------+
        |           | Pull inferencing |           | Pull model layer |
        +-----------| runtime (cuda)   |---------->| granite3-moe     |
                    +------------------+           +------------------+
						   | Repo options:    |
						   +-+-------+------+-+
						     |       |      |
						     v       v      v
					     +---------+ +------+ +----------+
					     | Hugging | | OCI  | | Ollama   |
					     | Face    | |      | | Registry |
					     +-------+-+ +---+--+ +-+--------+
						     |       |      |
						     v       v      v
						   +------------------+
						   | Start with       |
						   | cuda runtime     |
						   | and              |
						   | granite3-moe     |
						   +------------------+

Regard this alpha, everything is under development, so expect breaking changes, luckily it’s easy to reset everything and re-install:

rm -rf /var/lib/ramalama # only required if running as root user
rm -rf $HOME/.local/share/ramalama

and install again.

This project wouldn’t be possible without the help of other projects like:

llama.cpp
whisper.cpp
vllm
podman
huggingface

so if you like this tool, give some of these repos a ⭐, and hey, give us a ⭐ too while you are at it.

Matrix

Open to contributors

Keep your files stored safely and securely with the SanDisk 2TB Extreme Portable SSD. With over 69,505 ratings and an impressive 4.6 out of 5 stars, this product has been purchased over 8K+ times in the past month. At only $129.99, this Amazon’s Choice product is a must-have for secure file storage.

Help keep private content private with the included password protection featuring 256-bit AES hardware encryption. Order now for just $129.99 on Amazon!

Help Power Techcratic’s Future – Scan To Support

If Techcratic’s content and insights have helped you, consider giving back by supporting the platform with crypto. Every contribution makes a difference, whether it’s for high-quality content, server maintenance, or future updates. Techcratic is constantly evolving, and your support helps drive that progress.

As a solo operator who wears all the hats, creating content, managing the tech, and running the site, your support allows me to stay focused on delivering valuable resources. Your support keeps everything running smoothly and enables me to continue creating the content you love. I’m deeply grateful for your support, it truly means the world to me! Thank you!