Environment Variables Reference — Unique AI Documentation
Environment Variables Reference
Complete reference of all environment variables that configure the Web Search tool at the platform level. All variables must be available in the assistants-core pod environment at startup.
Variables are loaded via Pydantic Settings from the process environment and optionally from a .env file.
Core Settings
| Variable | Type | Default | Description |
|---|---|---|---|
WEB_SEARCH_MODE |
v1 / v2 |
v2 |
Default search mode for new spaces. |
ACTIVE_SEARCH_ENGINES |
JSON array | `[ | |
| "google" | |||
| ]` | Which search engines are available in the space configuration UI. | ||
ACTIVE_INHOUSE_CRAWLERS |
JSON array | `[ | |
| "basic", | |||
| "crawl4ai" | |||
| ]` | Which built-in crawlers are available. | ||
DISABLE_WEB_SEARCH_TOOL |
string | "false" |
Feature flag to disable the tool entirely. |
Google Search API
| Variable | Type | Default | Description |
|---|---|---|---|
GOOGLE_SEARCH_API_KEY |
string | null |
API key for Google Custom Search JSON API. |
GOOGLE_SEARCH_API_ENDPOINT |
string | null |
API endpoint URL (e.g., <https://customsearch.googleapis.com/customsearch/v1). |
GOOGLE_SEARCH_ENGINE_ID |
string | null |
Custom Search Engine ID (cx parameter). |
Bing / Azure AI
| Variable | Type | Default | Description |
|---|---|---|---|
AZURE_AI_PROJECT_ENDPOINT |
string | null |
Azure AI project endpoint URL. |
AZURE_AI_BING_RESOURCE_CONNECTION_STRING |
string | null |
Bing resource connection string from Azure. |
AZURE_AI_BING_AGENT_MODEL |
string | -- (required) |
The deployed model name used by the Bing Foundry Agent in auto-provisioned mode. |
AZURE_AI_ASSISTANT_ID |
string | null |
Pre-configured Foundry Agent ID. When set, skips auto-provisioning and uses this agent as-is. |
Brave Search API
| Variable | Type | Default | Description |
|---|---|---|---|
BRAVE_SEARCH_API_KEY |
string | null |
API key for Brave Search. |
BRAVE_SEARCH_API_ENDPOINT |
string | null |
Brave Search API endpoint URL. |
Jina AI
| Variable | Type | Default | Description |
|---|---|---|---|
JINA_API_KEY |
string | null |
API key for Jina AI. Also enables Jina search and crawler. |
JINA_SEARCH_API_ENDPOINT |
string | <https://s.jina.ai/> |
Jina Search API endpoint. |
JINA_READER_API_ENDPOINT |
string | <https://r.jina.ai/> |
Jina Reader API endpoint (used by crawler). |
Tavily
| Variable | Type | Default | Description |
|---|---|---|---|
TAVILY_API_KEY |
string | null |
API key for Tavily. Also enables Tavily search and crawler. |
Firecrawl
| Variable | Type | Default | Description |
|---|---|---|---|
FIRECRAWL_API_KEY |
string | null |
API key for Firecrawl. Also enables Firecrawl search and crawler. |
VertexAI (Google Cloud)
| Variable | Type | Default | Description |
|---|---|---|---|
VERTEXAI_SERVICE_ACCOUNT_CREDENTIALS |
string | null |
GCP service account credentials (JSON string). |
VERTEXAI_SERVICE_ACCOUNT_SCOPES |
list of strings | null |
OAuth scopes for the service account. |
Custom API
| Variable | Type | Default | Description |
|---|---|---|---|
CUSTOM_WEB_SEARCH_API_METHOD |
GET / POST |
null |
HTTP method. When set, the field is hidden from space-level config. |
CUSTOM_WEB_SEARCH_API_ENDPOINT |
string | null |
API endpoint URL. When set, the field is hidden from space-level config. |
CUSTOM_WEB_SEARCH_API_HEADERS |
string (JSON) | null |
Request headers. When set, the field is hidden from space-level config. |
CUSTOM_WEB_SEARCH_API_ADDITIONAL_QUERY_PARAMS |
string (JSON) | null |
Additional query parameters. When set, the field is hidden from space-level config. |
CUSTOM_WEB_SEARCH_API_ADDITIONAL_BODY_PARAMS |
string (JSON) | null |
Additional request body parameters. When set, the field is hidden from space-level config. |
CUSTOM_WEB_SEARCH_API_CLIENT_CONFIG |
string (JSON) | null |
HTTP client configuration (e.g., SSL settings). |
Proxy Settings
| Variable | Type | Default | Description |
|---|---|---|---|
PROXY_AUTH_MODE |
none / username_password / ssl_tls |
none |
Proxy authentication mode. |
PROXY_PROTOCOL |
http / https |
http |
Proxy protocol. |
PROXY_HOST |
string | null |
Proxy hostname. |
PROXY_PORT |
integer | null |
Proxy port number. |
PROXY_HEADERS |
dict (JSON) | {} |
Additional headers sent through the proxy. |
PROXY_SSL_CA_BUNDLE_PATH |
string | null |
Custom CA bundle path for proxy SSL verification. |
PROXY_USERNAME |
string | null |
Username for username_password auth mode. |
PROXY_PASSWORD |
string | null |
Password for username_password auth mode. |
PROXY_SSL_CERT_PATH |
string | null |
Client certificate path for ssl_tls auth mode. |
PROXY_SSL_KEY_PATH |
string | null |
Client key path for ssl_tls auth mode. |
Azure Identity
| Variable | Type | Default | Description |
|---|---|---|---|
UNIQUE_PRIVATE_ENDPOINT_TRANSPORT_ENABLED |
boolean | false |
Whether to use Azure private endpoint transport. |
DEFAULT_AZURE_IDENTITY_CREDENTIAL_TYPE |
workload / default |
default |
Azure identity credential type. Use workload for AKS workload identity. |
DEFAULT_AZURE_IDENTITY_CREDENTIALS_VALIDATE_TOKEN_URL |
string | <https://management.azure.com/.default> |
URL used to validate Azure identity tokens. |
Other
| Variable | Type | Default | Description |
|---|---|---|---|
ENV |
string | null |
Environment name (e.g., production, staging). |
LOG_LEVEL |
string | null |
Logging level (e.g., INFO, DEBUG). |
TIKTOKEN_CACHE_DIR |
string | ./tiktoken_cache/ |
Directory for caching tokenizer data. |