tarzi 0.2.3

Rust-native lite search for AI applications
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

# Tarzi environment configuration
# Copy values into your shell / process environment (there is no tarzi.toml).
# Precedence: CLI parameters > environment variables > defaults

# -----------------------------------------------------------------------------
# General
# -----------------------------------------------------------------------------
# TARZI_LOG_LEVEL=info
# TARZI_TIMEOUT=30

# -----------------------------------------------------------------------------
# Fetcher
# -----------------------------------------------------------------------------
# TARZI_FETCHER_BROWSER=true            # true | false — allow headless browser as fetch fallback (default true)
# TARZI_FETCHER_FORMAT=markdown         # markdown | html | text | json | yaml
# TARZI_USER_AGENT=Mozilla/5.0 ...
# TARZI_FETCHER_TIMEOUT=30
# TARZI_PROXY=http://127.0.0.1:7890     # fallback; HTTP(S)_PROXY wins at use time
# TARZI_WEB_DRIVER=chromedriver         # chromedriver | geckodriver
# TARZI_WEB_DRIVER_URL=http://localhost:4444

# -----------------------------------------------------------------------------
# Search
# -----------------------------------------------------------------------------
# Single engine or comma-separated ordered failover list (default: duckduckgo,bing,brave):
# TARZI_SEARCH_ENGINE=bing              # bing | google | google_serper | brave | duckduckgo | baidu | sogou_weixin | tavily | googleai | searxng
# TARZI_SEARCH_ENGINE=duckduckgo,bing,brave
# TARZI_SEARCH_BROWSER=true             # true | false — allow browser as search fallback (default true)
# TARZI_SEARCH_LIMIT=5
# TARZI_QUERY_PATTERN=https://example.com/search?q={query}

# Search access cascade (always): API if credentials present → plain HTTP → headless browser (if enabled).
# Fetch access cascade (always): plain HTTP → headless browser (if TARZI_FETCHER_BROWSER enabled).
# API-only engines (google_serper, tavily, googleai, searxng) are skipped when their env key/host is missing.

# -----------------------------------------------------------------------------
# Standard / engine-specific API keys and hosts (not TARZI_* prefixed)
# -----------------------------------------------------------------------------
# HTTPS_PROXY=http://proxy.example.com:8080
# HTTP_PROXY=http://proxy.example.com:8080
# BRAVE_API_KEY=your-brave-api-key
# SERPER_API_KEY=your-serper-api-key
# TAVILY_API_KEY=your-tavily-api-key
# GEMINI_API_KEY=your-gemini-api-key
# SEARX_HOST=http://localhost:8080