python>= 3.11.xuv- package managerdocker
In order for the application to have everything it needs,
a few Environment Variables will need to be set in
a .env-file at the root of this project (./.env):
# --- LLM Settings ---
AZURE_API_KEY="YOUR_API_KEY_FOR_AZURE_OPENAI"
AZURE_API_BASE="https://yourdeploymentregion.api.cognitive.microsoft.com/"
AZURE_API_VERSION="2024-12-01-preview"
DEFAULT_MODEL="azure/<deployment-name>"
DEFAULT_MODEL_TEMPERATURE=1.0
DEFAULT_MODEL_MAX_TOKENS=8000
# --- Logging ---
LOG_LEVEL="WARNING"
LOGGING_DIR="log"
GITHUB_PERSONAL_ACCESS_TOKEN="YOUR PERSONAL GITHUB ACCESS TOKEN WITH PERMISSIONS TO MANAGE YOUR REPOSITORIES."
run: uv sync to install the necessary python dependencies.
This project depends on two services:
dockerKernel-Memoryby Microsoft for RAG-Features
Setting up docker should be straightforward.
However, the application expects a Kernel-Memory instance
to listen at http://localhost:9001/.
For that, we need to start a dedicated docker-container.
In ./kernel-memory-config/appsettings.Development.json
you can find the configuration for the kernel-memory service.
Here, you will need to adjust two sections:
"AzureOpenAIEmbedding": {
"Auth": "ApiKey",
"AzureIdentityAudience": null,
"Endpoint": "https://YOUR_DEPLOYMENT_ENDPOINT.api.cognitive.microsoft.com/",
"APIKey": "YOUR_API_KEY",
"Deployment": "YOUR_EMBEDDING_MODEL_DEPLOYMENT_NAME",
"MaxTokenTotal": 8191,
"Tokenizer": "cl100k",
"EmbeddingDimensions": null,
"MaxEmbeddingBatchSize": 1,
"MaxRetries": 10,
"TrustedCertificateThumbprints": [],
"APIType": "EmbeddingGeneration"
},
"AzureOpenAIText": {
"Auth": "ApiKey",
"AzureIdentityAudience": null,
"Endpoint": "https://YOUR_DEPLOYMENT_ENDPOINT.api.cognitive.microsoft.com/",
"APIKey": "YOUR_API_KEY",
"Deployment": "YOUR_TEXT_GENERATION_MODEL_DEPLOYMENT_NAME (e.g. gpt-4.1-nano)",
"MaxTokenTotal": 16384,
"Tokenizer": "o200k",
"APIType": "ChatCompletion",
"MaxRetries": 10,
"TrustedCertificateThumbprints": []
},Once you have that sorted out, you can start the docker-container for Kernel-Memory with:
docker run --volume ./kernel-memory-configs/appsettings.Development.json:/app/appsettings.Production.json -it --rm -p 9001:9001 kernelmemory/serviceAfter you have performed the necessary setup steps, just run:
uv run main.pyYou can find a demo-prompt in ./demo_prompt.txt that
will instruct your agents to implement the Task Management Application that is specified
in ./files_for_memory