Skip to content

Version 1.7 of the documentation is no longer actively maintained. The site that you are currently viewing is an archived snapshot. For up-to-date documentation, see the latest version.

Configuration Reference

HugeGraph-LLM reads runtime settings from hugegraph-llm/.env. Prompts are stored separately in hugegraph-llm/src/hugegraph_llm/resources/demo/config_prompt.yaml and are not written to .env.

Create or update the files from configuration-class defaults with:

cd hugegraph-ai/hugegraph-llm
python -m hugegraph_llm.config.generate --update

.env contains keys and passwords. Do not commit it to version control.

Basic Options

SettingDefaultDescription
LANGUAGEENPrompt language: EN or CN
CHAT_LLM_TYPEopenaiAnswer model: openai, litellm, or ollama/local
EXTRACT_LLM_TYPEopenaiInformation extraction model; same choices as above
TEXT2GQL_LLM_TYPEopenaiText2Gremlin model; same choices as above
EMBEDDING_TYPEopenaiEmbedding model; same choices as above, or empty
RERANKER_TYPEemptycohere or siliconflow
KEYWORD_EXTRACT_TYPEllmllm, textrank, or hybrid
WINDOW_SIZE3TextRank window size, from 1 to 10
HYBRID_LLM_WEIGHTS0.5Weight of LLM results in hybrid mode, from 0 to 1

OpenAI-Compatible APIs

Chat, extraction, and Text2Gremlin can use different endpoints, keys, and models.

PurposeAPI baseKeyModelDefault maximum tokens
AnswerOPENAI_CHAT_API_BASEOPENAI_CHAT_API_KEYOPENAI_CHAT_LANGUAGE_MODELOPENAI_CHAT_TOKENS=8192
ExtractionOPENAI_EXTRACT_API_BASEOPENAI_EXTRACT_API_KEYOPENAI_EXTRACT_LANGUAGE_MODELOPENAI_EXTRACT_TOKENS=256
Text2GremlinOPENAI_TEXT2GQL_API_BASEOPENAI_TEXT2GQL_API_KEYOPENAI_TEXT2GQL_LANGUAGE_MODELOPENAI_TEXT2GQL_TOKENS=4096
EmbeddingOPENAI_EMBEDDING_API_BASEOPENAI_EMBEDDING_API_KEYOPENAI_EMBEDDING_MODELNot applicable

The default API base is https://api.openai.com/v1. The default language model for all three tasks is gpt-4.1-mini, and the default embedding model is text-embedding-3-small.

OPENAI_BASE_URL and OPENAI_API_KEY provide general fallback values. Embeddings also support OPENAI_EMBEDDING_BASE_URL and OPENAI_EMBEDDING_API_KEY as fallback values.

LiteLLM

PurposeAPI baseKeyModelDefault maximum tokens
AnswerLITELLM_CHAT_API_BASELITELLM_CHAT_API_KEYLITELLM_CHAT_LANGUAGE_MODELLITELLM_CHAT_TOKENS=8192
ExtractionLITELLM_EXTRACT_API_BASELITELLM_EXTRACT_API_KEYLITELLM_EXTRACT_LANGUAGE_MODELLITELLM_EXTRACT_TOKENS=256
Text2GremlinLITELLM_TEXT2GQL_API_BASELITELLM_TEXT2GQL_API_KEYLITELLM_TEXT2GQL_LANGUAGE_MODELLITELLM_TEXT2GQL_TOKENS=4096
EmbeddingLITELLM_EMBEDDING_API_BASELITELLM_EMBEDDING_API_KEYLITELLM_EMBEDDING_MODELNot applicable

The default language model is openai/gpt-4.1-mini, and the default embedding model is openai/text-embedding-3-small. Model names generally use the provider/model form; supported values depend on the LiteLLM service.

Ollama

PurposeHostPortModel
AnswerOLLAMA_CHAT_HOSTOLLAMA_CHAT_PORTOLLAMA_CHAT_LANGUAGE_MODEL
ExtractionOLLAMA_EXTRACT_HOSTOLLAMA_EXTRACT_PORTOLLAMA_EXTRACT_LANGUAGE_MODEL
Text2GremlinOLLAMA_TEXT2GQL_HOSTOLLAMA_TEXT2GQL_PORTOLLAMA_TEXT2GQL_LANGUAGE_MODEL
EmbeddingOLLAMA_EMBEDDING_HOSTOLLAMA_EMBEDDING_PORTOLLAMA_EMBEDDING_MODEL

The default host is 127.0.0.1 and the default port is 11434. Model names have no defaults; pull the required models in Ollama before use.

Reranking

SettingDefaultDescription
COHERE_BASE_URLhttps://api.cohere.com/v1/rerankCohere rerank endpoint; CO_API_URL is a fallback
RERANKER_API_KEYemptyCohere or SiliconFlow key
RERANKER_MODELemptyModel name supported by the service

HugeGraph Connection and Retrieval Limits

SettingDefaultDescription
GRAPH_URL127.0.0.1:8080HugeGraph address; it is not split into IP and port
GRAPH_NAMEhugegraphGraph name
GRAPH_USERadminUser name
GRAPH_PWDxxxPassword
GRAPH_SPACEemptyGraphSpace name
LIMIT_PROPERTYFalseWhether to limit returned properties; read as a string by the configuration class
MAX_GRAPH_PATH10Maximum graph path length
MAX_GRAPH_ITEMS30Maximum number of graph retrieval items
EDGE_LIMIT_PRE_LABEL8Result limit for each edge label
VECTOR_DIS_THRESHOLD0.9Results beyond this vector-distance threshold are ignored
TOPK_PER_KEYWORD1Candidates per keyword
TOPK_RETURN_RESULTS20Results returned after reranking

External Vector Databases

The default implementation can use local FAISS. After enabling optional dependencies, the following settings are also available:

SettingDefault
QDRANT_HOSTempty
QDRANT_PORT6333
QDRANT_API_KEYempty
MILVUS_HOSTempty
MILVUS_PORT19530
MILVUS_USERempty
MILVUS_PASSWORDempty
cd hugegraph-ai
uv sync --package hugegraph-llm --extra vectordb

Login and Log API

SettingDefaultDescription
ENABLE_LOGINFalseWhether to require a Bearer token; read as a string by the configuration class
USER_TOKEN4321Token for the Web UI and regular APIs
ADMIN_TOKENxxxxAdministrator token used by /logs

/logs returns 403 when ADMIN_TOKEN is empty or still set to xxxx. Replace both the user and administrator tokens in production.

Minimal OpenAI Configuration

LANGUAGE=EN
CHAT_LLM_TYPE=openai
EXTRACT_LLM_TYPE=openai
TEXT2GQL_LLM_TYPE=openai
EMBEDDING_TYPE=openai

OPENAI_API_KEY=your-api-key
OPENAI_BASE_URL=https://api.openai.com/v1
OPENAI_CHAT_LANGUAGE_MODEL=gpt-4.1-mini
OPENAI_EXTRACT_LANGUAGE_MODEL=gpt-4.1-mini
OPENAI_TEXT2GQL_LANGUAGE_MODEL=gpt-4.1-mini
OPENAI_EMBEDDING_MODEL=text-embedding-3-small

GRAPH_URL=127.0.0.1:8080
GRAPH_NAME=hugegraph
GRAPH_USER=admin
GRAPH_PWD=your-password

Configuration Loading

Configuration classes supply code defaults and then apply overrides from .env and the process environment. The Web UI and configuration APIs can update current settings at runtime and write supported fields back to .env. Restart the service after editing .env manually; prompt YAML can be refreshed by the page-loading logic.

Configuration definitions are in:

  • hugegraph-llm/src/hugegraph_llm/config/llm_config.py
  • hugegraph-llm/src/hugegraph_llm/config/hugegraph_config.py
  • hugegraph-llm/src/hugegraph_llm/config/admin_config.py
  • hugegraph-llm/src/hugegraph_llm/config/prompt_config.py
Last modified 2026-09-04 · docs: sync 1.7 documentation (4211f768b)