1min-Relay

Relay 1min AI API to OpenAI Structure in 1 minute.

Don't forget to star this repository if you like it!

Features

• bolt.diy Support: Compatible with bolt.diy for seamless integration.
• Conversation History: Retain and manage conversation history effortlessly.
• Broad Client Compatibility: Works with most clients supporting OpenAI Custom Endpoint.
• Fast and Reliable Relay: Relays 1min AI API to OpenAI-compatible structure within 1 minute.
• User-Friendly: Easy to set up and quick to use.
• Model Exposure Control: Expose all or a predefined subset of 1min.ai-supported models.
• Streaming Support: Enables real-time streaming for faster interactions.
• Non-Streaming Support: Compatible with non-streaming workflows.
• Docker Support: Deploy easily with Docker for fast and consistent setup.
• Multi-Document Support: Supports document uploads for enhanced functionality. Some include .docx, .pdf, .txt, .yaml and etc.
• Image Support: Supports image uploads for enhanced functionality.
• ARM64 and AMD64 Support: Compatible with both ARM64 and AMD64 architectures.
• Multiple Requests: Supports multiple requests simultaneously.

Installation:

Clone the git repo into your machine with:

git clone https://github.com/kokofixcomputers/1min-relay.git

Bare-metal

To install dependencies, run:

pip install -r requirements.txt

and then run python3 main.py

Depending on your system, you may need to run python instead of python3.

Docker

Running 1min-relay in docker is the easiet method. Please note that the connection ip address displayed on server start will be wrong when in docker.

Pre-Built images

1. Pull the Docker Image:

docker pull kokofixcomputers/1min-relay:latest

2. To encrease security, 1min-relay will require it's own network to be able to communicate with memcached. To create a network, run:

docker network create 1min-relay-network

3. Run Memcached.

docker run -d --name memcached --network 1min-relay-network memcached

4. Run the 1min-relay Container:

docker run -d --name 1min-relay-container --network 1min-relay-network -p 5001:5001 \
  -e SUBSET_OF_ONE_MIN_PERMITTED_MODELS="mistral-nemo,gpt-4o-mini,deepseek-chat" \
  -e PERMIT_MODELS_FROM_SUBSET_ONLY=True \
  kokofixcomputers/1min-relay:latest

Environment Variables:

• SUBSET_OF_ONE_MIN_PERMITTED_MODELS: Specifies a subset of 1min.ai models to expose. Default: mistral-nemo,gpt-4o,deepseek-chat.
• PERMIT_MODELS_FROM_SUBSET_ONLY: Restricts model usage to the specified subset. Set to True to enforce this restriction or False to allow all models supported by 1min.ai. Default: True.

Self-Build

1. Build the Docker Image From the project directory (where Dockerfile and main.py reside), run:

docker build -t 1min-relay:latest .

2. To encrease security, 1min-relay will require it's own network to be able to communicate with memcached. To create a network, run:

docker network create 1min-relay-network

3. Run Memcached.

docker run -d --name memcached --network 1min-relay-network memcached

4. Run the 1min-relay Container:

docker run -d --name 1min-relay-container --network 1min-relay-network -p 5001:5001 \
  -e SUBSET_OF_ONE_MIN_PERMITTED_MODELS="mistral-nemo,gpt-4o-mini,deepseek-chat" \
  -e PERMIT_MODELS_FROM_SUBSET_ONLY=True \
  1min-relay:latest

• -d runs the container in detached (background) mode.
• -p 5001:5001 maps your host’s port 5001 to the container’s port 5001.
• --name 1min-relay-container is optional, but it makes it easier to stop or remove the container later.
• -e: Specifies environment variables.
• SUBSET_OF_ONE_MIN_PERMITTED_MODELS: Specifies a subset of 1min.ai models to expose. Default: mistral-nemo,gpt-4o,deepseek-chat.
• PERMIT_MODELS_FROM_SUBSET_ONLY: Restricts model usage to the specified subset. Set to True to enforce this restriction or False to allow all models supported by 1min.ai. Default: False.

4. Verify It’s Running Check logs (optional):

docker logs -f 1min-relay-container

You should see your Flask server output: “Server is ready to serve at …”

Test from your host machine:

curl -X GET http://localhost:5001/v1/models

5. Stopping or Removing the Container To stop the container:

docker stop 1min-relay-container

To remove the container:

docker rm 1min-relay-container

To remove the image entirely:

docker rmi 1min-relay:latest

Optional: Run with Docker Compose If you prefer Docker Compose, you can run the docker compose included with the repo:

Just run:

docker compose up -d

Compose will automatically do these things for you:

• Create a network
• Run Memcached
• Run the 1min-relay Container

Managing docker containers

You can also have multiple instances of 1min-relay running on the same machine with docker.

You just need to chznge the name parameter of the container so you can identify them.

To stop a 1min-relay container, run:

docker stop 1min-relay-container

To start it again, run:

docker start 1min-relay-container

Environment Variables

• SUBSET_OF_ONE_MIN_PERMITTED_MODELS: Specifies a subset of 1min.ai models to expose. Default: mistral-nemo,gpt-4o,deepseek-chat.
• PERMIT_MODELS_FROM_SUBSET_ONLY: Restricts model usage to the specified subset. Set to True to enforce this restriction or False to allow all models supported by 1min.ai. Default: False.