totem-llm 0.1.0

package/server/.env.example

@@ -0,0 +1,496 @@
+SERVER_PORT=3001
+JWT_SECRET="my-random-string-for-seeding" # Please generate random string at least 12 chars long.
+# JWT_EXPIRY="30d" # (optional) https://docs.anythingllm.com/configuration#custom-ttl-for-sessions
+SIG_KEY='passphrase' # Please generate random string at least 32 chars long.
+SIG_SALT='salt' # Please generate random string at least 32 chars long.
+
+# SQLite database path (relative to server/prisma/schema.prisma).
+# Change to a postgresql:// URL to use PostgreSQL instead.
+DATABASE_URL="file:../storage/anythingllm.db"
+
+###########################################
+######## LLM API SElECTION ################
+###########################################
+# LLM_PROVIDER='openai'
+# OPEN_AI_KEY=
+# OPEN_MODEL_PREF='gpt-4o'
+
+# LLM_PROVIDER='gemini'
+# GEMINI_API_KEY=
+# GEMINI_LLM_MODEL_PREF='gemini-2.0-flash-lite'
+
+# LLM_PROVIDER='azure'
+# AZURE_OPENAI_ENDPOINT=
+# AZURE_OPENAI_KEY=
+# AZURE_OPENAI_MODEL_PREF='my-gpt35-deployment' # This is the "deployment" on Azure you want to use. Not the base model.
+# EMBEDDING_MODEL_PREF='embedder-model' # This is the "deployment" on Azure you want to use for embeddings. Not the base model. Valid base model is text-embedding-ada-002
+
+# LLM_PROVIDER='anthropic'
+# ANTHROPIC_API_KEY=sk-ant-xxxx
+# ANTHROPIC_MODEL_PREF='claude-sonnet-4-6'
+# ANTHROPIC_CACHE_CONTROL="5m" # Enable prompt caching (5m=5min cache, 1h=1hour cache). Reduces costs and improves speed by caching system prompts.
+
+# LLM_PROVIDER='lmstudio'
+# LMSTUDIO_BASE_PATH='http://your-server:1234/v1'
+# LMSTUDIO_MODEL_PREF='Loaded from Chat UI' # this is a bug in LMStudio 0.2.17
+# LMSTUDIO_MODEL_TOKEN_LIMIT=4096
+# LMSTUDIO_AUTH_TOKEN='your-lmstudio-auth-token-here'
+
+# LLM_PROVIDER='localai'
+# LOCAL_AI_BASE_PATH='http://localhost:8080/v1'
+# LOCAL_AI_MODEL_PREF='luna-ai-llama2'
+# LOCAL_AI_MODEL_TOKEN_LIMIT=4096
+# LOCAL_AI_API_KEY="sk-123abc"
+
+# LLM_PROVIDER='ollama'
+# OLLAMA_BASE_PATH='http://host.docker.internal:11434'
+# OLLAMA_MODEL_PREF='llama2'
+# OLLAMA_MODEL_TOKEN_LIMIT=4096
+# OLLAMA_AUTH_TOKEN='your-ollama-auth-token-here (optional, only for ollama running behind auth - Bearer token)'
+# OLLAMA_RESPONSE_TIMEOUT=7200000 (optional, max timeout in milliseconds for ollama response to conclude. Default is 5min before aborting)
+
+# LLM_PROVIDER='togetherai'
+# TOGETHER_AI_API_KEY='my-together-ai-key'
+# TOGETHER_AI_MODEL_PREF='mistralai/Mixtral-8x7B-Instruct-v0.1'
+
+# LLM_PROVIDER='fireworksai'
+# FIREWORKS_AI_LLM_API_KEY='my-fireworks-ai-key'
+# FIREWORKS_AI_LLM_MODEL_PREF='accounts/fireworks/models/llama-v3p1-8b-instruct'
+
+# LLM_PROVIDER='perplexity'
+# PERPLEXITY_API_KEY='my-perplexity-key'
+# PERPLEXITY_MODEL_PREF='codellama-34b-instruct'
+
+# LLM_PROVIDER='deepseek'
+# DEEPSEEK_API_KEY=YOUR_API_KEY
+# DEEPSEEK_MODEL_PREF='deepseek-chat'
+
+# LLM_PROVIDER='openrouter'
+# OPENROUTER_API_KEY='my-openrouter-key'
+# OPENROUTER_MODEL_PREF='openrouter/auto'
+
+# LLM_PROVIDER='mistral'
+# MISTRAL_API_KEY='example-mistral-ai-api-key'
+# MISTRAL_MODEL_PREF='mistral-tiny'
+
+# LLM_PROVIDER='huggingface'
+# HUGGING_FACE_LLM_ENDPOINT=https://uuid-here.us-east-1.aws.endpoints.huggingface.cloud
+# HUGGING_FACE_LLM_API_KEY=hf_xxxxxx
+# HUGGING_FACE_LLM_TOKEN_LIMIT=8000
+
+# LLM_PROVIDER='groq'
+# GROQ_API_KEY=gsk_abcxyz
+# GROQ_MODEL_PREF=llama3-8b-8192
+
+# LLM_PROVIDER='koboldcpp'
+# KOBOLD_CPP_BASE_PATH='http://127.0.0.1:5000/v1'
+# KOBOLD_CPP_MODEL_PREF='koboldcpp/codellama-7b-instruct.Q4_K_S'
+# KOBOLD_CPP_MODEL_TOKEN_LIMIT=4096
+# KOBOLD_CPP_MAX_TOKENS=2048
+
+# LLM_PROVIDER='textgenwebui'
+# TEXT_GEN_WEB_UI_BASE_PATH='http://127.0.0.1:5000/v1'
+# TEXT_GEN_WEB_UI_TOKEN_LIMIT=4096
+# TEXT_GEN_WEB_UI_API_KEY='sk-123abc'
+
+# LLM_PROVIDER='generic-openai'
+# GENERIC_OPEN_AI_BASE_PATH='http://proxy.url.openai.com/v1'
+# GENERIC_OPEN_AI_MODEL_PREF='gpt-3.5-turbo'
+# GENERIC_OPEN_AI_MODEL_TOKEN_LIMIT=4096
+# GENERIC_OPEN_AI_API_KEY=sk-123abc
+# GENERIC_OPEN_AI_CUSTOM_HEADERS="X-Custom-Auth:my-secret-key,X-Custom-Header:my-value" (useful if using a proxy that requires authentication or other headers)
+
+# LLM_PROVIDER='litellm'
+# LITE_LLM_MODEL_PREF='gpt-3.5-turbo'
+# LITE_LLM_MODEL_TOKEN_LIMIT=4096
+# LITE_LLM_BASE_PATH='http://127.0.0.1:4000'
+# LITE_LLM_API_KEY='sk-123abc'
+
+# LLM_PROVIDER='novita'
+# NOVITA_LLM_API_KEY='your-novita-api-key-here' check on https://novita.ai/settings#key-management
+# NOVITA_LLM_MODEL_PREF='deepseek/deepseek-r1'
+
+# LLM_PROVIDER='cohere'
+# COHERE_API_KEY=
+# COHERE_MODEL_PREF='command-r'
+
+# LLM_PROVIDER='cometapi'
+# COMETAPI_LLM_API_KEY='your-cometapi-key-here' # Get one at https://api.cometapi.com/console/token
+# COMETAPI_LLM_MODEL_PREF='gpt-5-mini'
+# COMETAPI_LLM_TIMEOUT_MS=500 # Optional; stream idle timeout in ms (min 500ms)
+
+
+# LLM_PROVIDER='bedrock'
+# AWS_BEDROCK_LLM_ACCESS_KEY_ID=
+# AWS_BEDROCK_LLM_ACCESS_KEY=
+# AWS_BEDROCK_LLM_REGION=us-west-2
+# AWS_BEDROCK_LLM_MODEL_PREFERENCE=meta.llama3-1-8b-instruct-v1:0
+# AWS_BEDROCK_LLM_MODEL_TOKEN_LIMIT=8191
+# AWS_BEDROCK_LLM_CONNECTION_METHOD=iam
+# AWS_BEDROCK_LLM_MAX_OUTPUT_TOKENS=4096
+# AWS_BEDROCK_LLM_SESSION_TOKEN= # Only required if CONNECTION_METHOD is 'sessionToken'
+# or even use Short and Long Term API keys
+# AWS_BEDROCK_LLM_CONNECTION_METHOD="apiKey"
+# AWS_BEDROCK_LLM_API_KEY=
+
+# LLM_PROVIDER='apipie'
+# APIPIE_LLM_API_KEY='sk-123abc'
+# APIPIE_LLM_MODEL_PREF='openrouter/llama-3.1-8b-instruct'
+
+# LLM_PROVIDER='xai'
+# XAI_LLM_API_KEY='xai-your-api-key-here'
+# XAI_LLM_MODEL_PREF='grok-beta'
+
+# LLM_PROVIDER='zai'
+# ZAI_API_KEY="your-zai-api-key-here"
+# ZAI_MODEL_PREF="glm-4.5"
+
+# LLM_PROVIDER='nvidia-nim'
+# NVIDIA_NIM_LLM_BASE_PATH='http://127.0.0.1:8000'
+# NVIDIA_NIM_LLM_MODEL_PREF='meta/llama-3.2-3b-instruct'
+
+# LLM_PROVIDER='ppio'
+# PPIO_API_KEY='your-ppio-api-key-here'
+# PPIO_MODEL_PREF='deepseek/deepseek-v3/community'
+
+# LLM_PROVIDER='moonshotai'
+# MOONSHOT_AI_API_KEY='your-moonshot-api-key-here'
+# MOONSHOT_AI_MODEL_PREF='moonshot-v1-32k'
+
+# LLM_PROVIDER='foundry'
+# FOUNDRY_BASE_PATH='http://127.0.0.1:55776'
+# FOUNDRY_MODEL_PREF='phi-3.5-mini'
+# FOUNDRY_MODEL_TOKEN_LIMIT=4096
+
+# LLM_PROVIDER='giteeai'
+# GITEE_AI_API_KEY=
+# GITEE_AI_MODEL_PREF=
+# GITEE_AI_MODEL_TOKEN_LIMIT=
+
+# LLM_PROVIDER='docker-model-runner'
+# DOCKER_MODEL_RUNNER_BASE_PATH='http://127.0.0.1:12434'
+# DOCKER_MODEL_RUNNER_LLM_MODEL_PREF='phi-3.5-mini'
+# DOCKER_MODEL_RUNNER_LLM_MODEL_TOKEN_LIMIT=4096
+
+# LLM_PROVIDER='privatemode'
+# PRIVATEMODE_LLM_BASE_PATH='http://127.0.0.1:8080'
+# PRIVATEMODE_LLM_MODEL_PREF='gemma-3-27b'
+
+# LLM_PROVIDER='sambanova'
+# SAMBANOVA_LLM_API_KEY='xxx-xxx-xxx'
+# SAMBANOVA_LLM_MODEL_PREF='gpt-oss-120b'
+
+# LLM_PROVIDER='lemonade'
+# LEMONADE_LLM_BASE_PATH='http://127.0.0.1:8000'
+# LEMONADE_LLM_MODEL_PREF='Llama-3.2-1B-Instruct-GGUF'
+# LEMONADE_LLM_MODEL_TOKEN_LIMIT=8192
+# LEMONADE_LLM_API_KEY=
+
+# LLM_PROVIDER='minimax'
+# MINIMAX_API_KEY='sk-cp-...'
+# MINIMAX_MODEL_PREF='MiniMax-M2.7'
+
+# LLM_PROVIDER='cerebras'
+# CEREBRAS_API_KEY='your-cerebras-api-key-here'
+# CEREBRAS_MODEL_PREF='gpt-oss-120b'
+
+# LLM_PROVIDER='anythingllm-router'
+# MODEL_ROUTER_ID=1
+
+###########################################
+######## Embedding API SElECTION ##########
+###########################################
+# This will be the assumed default embedding seleciton and model
+# EMBEDDING_ENGINE='native'
+# EMBEDDING_MODEL_PREF='Xenova/all-MiniLM-L6-v2'
+
+# Only used if you are using an LLM that does not natively support embedding (openai or Azure)
+# EMBEDDING_ENGINE='openai'
+# OPEN_AI_KEY=sk-xxxx
+# EMBEDDING_MODEL_PREF='text-embedding-ada-002'
+
+# EMBEDDING_ENGINE='azure'
+# AZURE_OPENAI_ENDPOINT=
+# AZURE_OPENAI_KEY=
+# EMBEDDING_MODEL_PREF='my-embedder-model' # This is the "deployment" on Azure you want to use for embeddings. Not the base model. Valid base model is text-embedding-ada-002
+
+# EMBEDDING_ENGINE='localai'
+# EMBEDDING_BASE_PATH='http://localhost:8080/v1'
+# EMBEDDING_MODEL_PREF='text-embedding-ada-002'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=1000 # The max chunk size in chars a string to embed can be
+
+# EMBEDDING_ENGINE='ollama'
+# EMBEDDING_BASE_PATH='http://127.0.0.1:11434'
+# EMBEDDING_MODEL_PREF='nomic-embed-text:latest'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=8192
+
+# EMBEDDING_ENGINE='lmstudio'
+# EMBEDDING_BASE_PATH='https://localhost:1234/v1'
+# EMBEDDING_MODEL_PREF='nomic-ai/nomic-embed-text-v1.5-GGUF/nomic-embed-text-v1.5.Q4_0.gguf'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=8192
+
+# EMBEDDING_ENGINE='cohere'
+# COHERE_API_KEY=
+# EMBEDDING_MODEL_PREF='embed-english-v3.0'
+
+# EMBEDDING_ENGINE='voyageai'
+# VOYAGEAI_API_KEY=
+# EMBEDDING_MODEL_PREF='voyage-large-2-instruct'
+
+# EMBEDDING_ENGINE='litellm'
+# EMBEDDING_MODEL_PREF='text-embedding-ada-002'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=8192
+# LITE_LLM_BASE_PATH='http://127.0.0.1:4000'
+# LITE_LLM_API_KEY='sk-123abc'
+
+# EMBEDDING_ENGINE='generic-openai'
+# EMBEDDING_MODEL_PREF='text-embedding-ada-002'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=8192
+# EMBEDDING_BASE_PATH='http://127.0.0.1:4000'
+# GENERIC_OPEN_AI_EMBEDDING_API_KEY='sk-123abc'
+# GENERIC_OPEN_AI_EMBEDDING_MAX_CONCURRENT_CHUNKS=500
+# GENERIC_OPEN_AI_EMBEDDING_API_DELAY_MS=1000
+
+# EMBEDDING_ENGINE='gemini'
+# GEMINI_EMBEDDING_API_KEY=
+# EMBEDDING_MODEL_PREF='text-embedding-004'
+
+# EMBEDDING_ENGINE='openrouter'
+# EMBEDDING_MODEL_PREF='baai/bge-m3'
+# OPENROUTER_API_KEY=''
+
+# EMBEDDING_ENGINE='lemonade'
+# EMBEDDING_BASE_PATH='http://127.0.0.1:8000'
+# EMBEDDING_MODEL_PREF='Qwen3-embedder'
+# EMBEDDING_MODEL_MAX_CHUNK_LENGTH=8192
+
+###########################################
+######## Vector Database Selection ########
+###########################################
+# Enable all below if you are using vector database: Chroma.
+# VECTOR_DB="chroma"
+# CHROMA_ENDPOINT='http://localhost:8000'
+# CHROMA_API_HEADER="X-Api-Key"
+# CHROMA_API_KEY="sk-123abc"
+
+# Enable all below if you are using vector database: Chroma Cloud.
+# VECTOR_DB="chromacloud"
+# CHROMACLOUD_API_KEY="ck-your-api-key"
+# CHROMACLOUD_TENANT=
+# CHROMACLOUD_DATABASE=
+
+# Enable all below if you are using vector database: Pinecone.
+# VECTOR_DB="pinecone"
+# PINECONE_API_KEY=
+# PINECONE_INDEX=
+
+# Enable all below if you are using vector database: Astra DB.
+# VECTOR_DB="astra"
+# ASTRA_DB_APPLICATION_TOKEN=
+# ASTRA_DB_ENDPOINT=
+
+# Enable all below if you are using vector database: LanceDB.
+VECTOR_DB="lancedb"
+
+# Enable all below if you are using vector database: PG Vector.
+# VECTOR_DB="pgvector"
+# PGVECTOR_CONNECTION_STRING="postgresql://dbuser:dbuserpass@localhost:5432/yourdb"
+# PGVECTOR_TABLE_NAME="anythingllm_vectors" # optional, but can be defined
+
+# Enable all below if you are using vector database: Weaviate.
+# VECTOR_DB="weaviate"
+# WEAVIATE_ENDPOINT="http://localhost:8080"
+# WEAVIATE_API_KEY=
+
+# Enable all below if you are using vector database: Qdrant.
+# VECTOR_DB="qdrant"
+# QDRANT_ENDPOINT="http://localhost:6333"
+# QDRANT_API_KEY=
+
+# Enable all below if you are using vector database: Milvus.
+# VECTOR_DB="milvus"
+# MILVUS_ADDRESS="http://localhost:19530"
+# MILVUS_USERNAME=
+# MILVUS_PASSWORD=
+
+# Enable all below if you are using vector database: Zilliz Cloud.
+# VECTOR_DB="zilliz"
+# ZILLIZ_ENDPOINT="https://sample.api.gcp-us-west1.zillizcloud.com"
+# ZILLIZ_API_TOKEN=api-token-here
+
+###########################################
+######## Audio Model Selection ############
+###########################################
+# (default) use built-in whisper-small model.
+WHISPER_PROVIDER="local"
+
+# use openai hosted whisper model.
+# WHISPER_PROVIDER="openai"
+# OPEN_AI_KEY=sk-xxxxxxxx
+
+###########################################
+######## TTS/STT Model Selection ##########
+###########################################
+TTS_PROVIDER="native"
+
+# TTS_PROVIDER="openai"
+# TTS_OPEN_AI_KEY=sk-example
+# TTS_OPEN_AI_VOICE_MODEL=nova
+
+# TTS_PROVIDER="elevenlabs"
+# TTS_ELEVEN_LABS_KEY=
+# TTS_ELEVEN_LABS_VOICE_MODEL=21m00Tcm4TlvDq8ikWAM # Rachel
+
+# TTS_PROVIDER="generic-openai"
+# TTS_OPEN_AI_COMPATIBLE_KEY=sk-example
+# TTS_OPEN_AI_COMPATIBLE_MODEL=tts-1
+# TTS_OPEN_AI_COMPATIBLE_VOICE_MODEL=nova
+# TTS_OPEN_AI_COMPATIBLE_ENDPOINT="https://api.openai.com/v1"
+
+# CLOUD DEPLOYMENT VARIRABLES ONLY
+# AUTH_TOKEN="hunter2" # This is the password to your application if remote hosting.
+# STORAGE_DIR= # absolute filesystem path with no trailing slash
+
+###########################################
+######## PASSWORD COMPLEXITY ##############
+###########################################
+# Enforce a password schema for your organization users.
+# Documentation on how to use https://github.com/kamronbatman/joi-password-complexity
+#PASSWORDMINCHAR=8
+#PASSWORDMAXCHAR=250
+#PASSWORDLOWERCASE=1
+#PASSWORDUPPERCASE=1
+#PASSWORDNUMERIC=1
+#PASSWORDSYMBOL=1
+#PASSWORDREQUIREMENTS=4
+
+###########################################
+######## ENABLE HTTPS SERVER ##############
+###########################################
+# By enabling this and providing the path/filename for the key and cert,
+# the server will use HTTPS instead of HTTP.
+#ENABLE_HTTPS="true"
+#HTTPS_CERT_PATH="sslcert/cert.pem"
+#HTTPS_KEY_PATH="sslcert/key.pem"
+
+###########################################
+######## AGENT SERVICE KEYS ###############
+###########################################
+
+#------ SEARCH ENGINES -------
+#=============================
+#------ Google Search -------- https://programmablesearchengine.google.com/controlpanel/create
+# AGENT_GSE_KEY=
+# AGENT_GSE_CTX=
+
+#------ SerpApi ----------- https://serpapi.com/
+# AGENT_SERPAPI_API_KEY=
+# AGENT_SERPAPI_ENGINE=google
+
+#------ SearchApi.io ----------- https://www.searchapi.io/
+# AGENT_SEARCHAPI_API_KEY=
+# AGENT_SEARCHAPI_ENGINE=google
+
+#------ Serper.dev ----------- https://serper.dev/
+# AGENT_SERPER_DEV_KEY=
+
+#------ Bing Search ----------- https://portal.azure.com/
+# AGENT_BING_SEARCH_API_KEY=
+
+#------ Baidu Search ----------- https://cloud.baidu.com/doc/qianfan-api/s/Wmbq4z7e5
+# AGENT_BAIDU_SEARCH_API_KEY=
+
+#------ Serply.io ----------- https://serply.io/
+# AGENT_SERPLY_API_KEY=
+
+#------ SearXNG ----------- https://github.com/searxng/searxng
+# AGENT_SEARXNG_API_URL=
+
+#------ Tavily ----------- https://www.tavily.com/
+# AGENT_TAVILY_API_KEY=
+
+#------ Exa Search ----------- https://www.exa.ai/
+# AGENT_EXA_API_KEY=
+
+#------ Perplexity Search ----------- [https://console.perplexity.ai](https://console.perplexity.ai)
+# AGENT_PERPLEXITY_API_KEY=
+
+###########################################
+######## Other Configurations ############
+###########################################
+
+# Disable viewing chat history from the UI and frontend APIs.
+# See https://docs.anythingllm.com/configuration#disable-view-chat-history for more information.
+# DISABLE_VIEW_CHAT_HISTORY=1
+
+# Disable workspace deletion from the UI and APIs when this ENV is present with any value.
+# WORKSPACE_DELETION_PROTECTION=1
+
+# Enable simple SSO passthrough to pre-authenticate users from a third party service.
+# See https://docs.anythingllm.com/configuration#simple-sso-passthrough for more information.
+# SIMPLE_SSO_ENABLED=1
+# SIMPLE_SSO_NO_LOGIN=1
+# SIMPLE_SSO_NO_LOGIN_REDIRECT=https://your-custom-login-url.com (optional)
+
+# Allow scraping of any IP address in collector - must be string "true" to be enabled
+# See https://docs.anythingllm.com/configuration#local-ip-address-scraping for more information.
+# COLLECTOR_ALLOW_ANY_IP="true"
+
+# Port the collector listens on. Must match the collector process when running services separately.
+# COLLECTOR_PORT=8888
+
+# Specify the target languages for when using OCR to parse images and PDFs.
+# This is a comma separated list of language codes as a string. Unsupported languages will be ignored.
+# Default is English. See https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html for a list of valid language codes.
+# TARGET_OCR_LANG=eng,deu,ita,spa,fra,por,rus,nld,tur,hun,pol,ita,spa,fra,por,rus,nld,tur,hun,pol
+
+# Runtime flags for built-in pupeeteer Chromium instance
+# This is only required on Linux machines running AnythingLLM via Docker
+# and do not want to use the --cap-add=SYS_ADMIN docker argument
+# ANYTHINGLLM_CHROMIUM_ARGS="--no-sandbox,--disable-setuid-sandbox"
+
+# This enables HTTP request/response logging in development. Set value to a truthy string to enable, leave empty value or comment out to disable.
+# ENABLE_HTTP_LOGGER=""
+# This enables timestamps for the HTTP Logger. Set value to a truthy string to enable, leave empty value or comment out to disable.
+# ENABLE_HTTP_LOGGER_TIMESTAMPS=""
+
+# Disable Swagger API documentation endpoint.
+# Set to "true" to disable the /api/docs endpoint (recommended for production deployments).
+# DISABLE_SWAGGER_DOCS="true"
+
+# Disable MCP cooldown timer for agent calls
+# this can lead to infinite recursive calls of the same function
+# for some model/provider combinations
+# MCP_NO_COOLDOWN="true
+
+# Allow native tool calling for specific providers.
+# This can VASTLY improve performance and speed of agent calls.
+# Check code for supported providers who can be enabled here via this flag
+# PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING="generic-openai,bedrock,localai,groq,litellm,openrouter"
+
+# (optional) Maximum number of tools an agent can chain for a single response.
+# This prevents some lower-end models from infinite recursive tool calls.
+# AGENT_MAX_TOOL_CALLS=10
+
+# Enable agent tool reranking to reduce token usage by selecting only the most relevant tools
+# for each query. Uses the native embedding reranker to score tools against the user's prompt.
+# Set to "true" to enable. This can reduce token costs by 80% when you have
+# many tools/MCP servers enabled.
+# AGENT_SKILL_RERANKER_ENABLED="true"
+# AGENT_SKILL_RERANKER_TOP_N=15 # (optional) Number of top tools to keep after reranking (default: 15)
+
+# (optional) Memory extraction background job settings.
+# MEMORY_EXTRACTION_INTERVAL="15m" # How often the extraction job runs (default: 15m)
+# MEMORY_IDLE_THRESHOLD_MS=1200000 # Min ms since last chat before extraction runs (default: 1200000 / 20min). Set to 0 to disable the idle check.
+
+# (optional) Maximum number of scheduled jobs that can run concurrently.
+# Default is 1. Increase if using a cloud LLM provider with high rate limits.
+# SCHEDULED_JOB_MAX_CONCURRENT=1
+
+# (optional) Maximum time in milliseconds a scheduled job can run before being terminated.
+# Default is 5 minutes (300000ms).
+# SCHEDULED_JOB_TIMEOUT_MS=300000
+
+# (optional) Comma-separated list of skills that are auto-approved.
+# This will allow the skill to be invoked without user interaction.
+# AGENT_AUTO_APPROVED_SKILLS=create-pdf-file,create-word-file

package/server/.flowconfig

@@ -0,0 +1,30 @@
+# How to config: https://flow.org/en/docs/config/
+[version]
+
+[options]
+all=false
+emoji=false
+include_warnings=false
+lazy_mode=false
+
+[include]
+
+[ignore]
+.*/node_modules/resolve/test/.*
+
+[untyped]
+# <PROJECT_ROOT>/src/models/.*
+
+[declarations]
+
+[libs]
+
+[lints]
+all=warn
+
+[strict]
+nonstrict-import
+unclear-type
+unsafe-getters-setters
+untyped-import
+untyped-type-import

package/server/.nvmrc

@@ -0,0 +1 @@
+v18.18.0