Minipilot is a highly configurable server. Read the environment variables to customize the server's behavior.
| Variable | Description | Default |
|---|---|---|
OPENAI_API_KEY |
This is the OpenAI token. | |
OPENAI_MODEL |
You can specify the desired LLM. | gpt-4o-mini |
DB_SERVICE |
Redis IP or hostname. | 127.0.0.1 |
DB_PORT |
Redis port. | 6379 |
DB_PWD |
Redis password. | |
MINIPILOT_DEBUG |
It enables the debugging mode. | True |
MINIPILOT_LOG |
Where the log file will be saved. | ./minipilot.log |
MINIPILOT_ENDPOINT |
The UI invokes the local REST API, provide an endpoint like http://localhost:5005. |
|
MINIPILOT_HISTORY_TIMEOUT |
Time to Live of the entries in the conversation history. | 604800 |
MINIPILOT_HISTORY_LENGTH |
The conversation history is limited, when the max configured number of interactions is reached, the older entries are removed. | 30 |
MINIPILOT_RATE_LIMITER_ENABLED |
Whether the rate limiter is enabled. | True |
MINIPILOT_RATE_LIMITER_CRITERIA |
Limiting the rate based on ip , session, or all. |
session |
MINIPILOT_RATE_LIMITER_ALLOW |
Maximum operations per minute allowed by the rate limiter. | 10 |
MINIPILOT_CONVERSATION_LENGTH |
Maximum size of the minipilot:conversation stream storing all the interactions. |
10000 |
MINIPILOT_LOG_LENGTH |
Maximum size of the minipilot:log stream storing the application log. |
1000 |
MINIPILOT_CONTEXT_LENGTH |
Number of entries retrieved from the database for RAG. | 5 |
MINIPILOT_CONTEXT_RELEVANCE_SCORE |
Threshold to limit the returned entries retrieved from the database for RAG. | 0.78 |
MINIPILOT_LLM_TIMEOUT |
Timeout to control eventual LLM slowness. | 10 |
MINIPILOT_CACHE_TTL |
Time to Live of the entries in the semantic cache. | 3600 * 24 * 30 |
MINIPILOT_CACHE_THRESHOLD |
Semantic similarity threshold for the results retrieved from the semantic cache. | 0.1 |
MINIPILOT_CACHE_ENABLED |
Whether the semantic cache is enabled. | True |