This project allows you to customize the Prompt for each starter, optimizes the UI with CSS, and provides a Docker image for easy deployment. Simply configure the config.json for a ready-to-use setup. Below is a guide on how to set up and run this project.
-
Install
uvcurl -LsSf https://astral.sh/uv/install.sh | sh -
Configure config.toml
[MODEL_MAP]
[[MODEL_MAP]."{name}"] # Corresponds to a model
profile_name = "Model A" # Display name of the model
model_name = "model-a" # Actual model name
markdown_description = "Description of Model A" # Model description
icon = "/public/model/model-a.svg" # Model icon
# Starter settings, up to four starters
[STARTERS]
START_1_LABEL = "Professional Copywriting Assistant" # Content displayed in the UI for the starter
START_1_MESSAGE = "Help me write a copy! Master" # Message sent when the starter is launched
START_1_ICON = "/public/icons/copywriter.webp" # Icon for the starter
# One prompt corresponds to one starter, the first prompt_1 is the default prompt
# From the second prompt_2 corresponds to the first starter prompt, must match the number of starters
# Default prompt: (prompt_1); starter: (prompt_2,prompt_3,prompt_4,prompt_5) up to four
[PROMPTS]
PROMPT_1 = """
xxxxxx
"""
-
Sync Environment
uv sync
-
Run the Project
cd src uv run -- chainlit run app.py -w --host 0.0.0.0 --port 8000- The
--hostand--portparameters are optional and can be adjusted as needed. Refer to chainlit for details.
- The
mv config.toml.exp config.toml
# Modify .env as needed
docker run -d -p 8000:8000 -v ./config.toml:/app/config.toml bitfennec/llmer:latestVisit http://localhost:8000 to view the running application.
git clone https://github.com/Mgrsc/llmer.git
cd llmerEdit docker-compose.yaml as needed
services: llmer: image: bitfennec/llmer:latest container_name: llmer ports: - 30004:8000 volumes: - ./config.toml:/app/config.toml # - ./src/public/model:/app/src/public/model # Optional, if you need to add icons
docker-compose up -d
Visit http://localhost:8000 to view the running application.
If you have any suggestions for improvements or want to report issues, feel free to submit an issue or pull request (PR).
This project is based on Chainlit and uses the Apache 2.0 license. For details, please refer to the LICENSE file.