-
Notifications
You must be signed in to change notification settings - Fork 23
Expand file tree
/
Copy path.env.example
More file actions
111 lines (99 loc) · 3.88 KB
/
Copy path.env.example
File metadata and controls
111 lines (99 loc) · 3.88 KB
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
# AF Deep Research - Environment Configuration
# ================================================
# Copy this file to .env and fill in your API keys:
# cp .env.example .env
#
# Then run:
# docker-compose up
# ===================
# Required API Keys
# ===================
# OpenRouter API Key (for LLM inference)
# Get yours at: https://openrouter.ai/keys
OPENROUTER_API_KEY=sk-or-your-openrouter-api-key-here
# ===================
# Web Search API Keys
# ===================
# Configure at least ONE of the following search providers.
# The system will auto-detect which one is available.
# Jina AI (recommended - free tier available)
# Get yours at: https://jina.ai/
JINA_API_KEY=jina_your-jina-api-key-here
# Tavily (AI-native search, 1000 free/month)
# Get yours at: https://tavily.com/
# TAVILY_API_KEY=tvly-your-tavily-api-key-here
# Firecrawl (search + scraping, 500 free credits)
# Get yours at: https://firecrawl.dev/
# FIRECRAWL_API_KEY=fc-your-firecrawl-api-key-here
# Serper (Google SERP, 2500 free/month)
# Get yours at: https://serper.dev/
# SERPER_API_KEY=your-serper-api-key-here
# ===================
# Optional Settings
# ===================
# Force a specific search provider (auto-detect if not set)
# Options: jina, tavily, firecrawl, serper
# SEARCH_PROVIDER=jina
# ===================
# LLM Model Configuration
# ===================
# Default LLM model to use (via OpenRouter)
# Format: openrouter/<provider>/<model-name>
# See all models: https://openrouter.ai/models
#
# RECOMMENDED MODELS (all verified to work with complex JSON output):
#
# --- Best Value (Default) ---
# openrouter/deepseek/deepseek-chat-v3.1 # $0.15/$0.75 per 1M tokens - excellent reasoning, 128K context
# openrouter/deepseek/deepseek-v3.2 # Latest DeepSeek, similar pricing
#
# --- Premium (Highest Quality) ---
# openrouter/anthropic/claude-sonnet-4 # $3/$15 per 1M tokens - excellent for complex analysis
# openrouter/anthropic/claude-opus-4 # $15/$75 per 1M tokens - best reasoning capability
#
# --- Budget-Friendly ---
# openrouter/google/gemini-2.5-flash # $0.30/$2.50 per 1M tokens - fast, 1M context
# openrouter/google/gemini-2.5-flash-lite # Even cheaper, optimized for speed
#
# --- Open Source ---
# openrouter/qwen/qwen2.5-72b-instruct # Excellent JSON/structured output
# openrouter/qwen/qwen3-235b-a22b # Latest Qwen, very capable
# openrouter/meta-llama/llama-3.3-70b-instruct # Good general purpose
# openrouter/meta-llama/llama-3.1-405b-instruct # Largest open model
#
DEFAULT_MODEL=openrouter/deepseek/deepseek-chat-v3.1
# LLM temperature (0.0-1.0, lower = more deterministic)
TEMPERATURE=0.6
# ===================
# Local Ollama Configuration (Optional)
# ===================
# To use a local Ollama deployment instead of OpenRouter:
# 1. Set OLLAMA_BASE_URL to your Ollama server
# 2. Change DEFAULT_MODEL to use ollama/<model-name> format
# 3. OPENROUTER_API_KEY is not required when using Ollama
#
# IMPORTANT: When running via Docker, use host.docker.internal instead of localhost
# to reach Ollama running on your host machine:
#
# OLLAMA_BASE_URL=http://host.docker.internal:11434
# DEFAULT_MODEL=ollama/llama3.2
#
# For non-Docker (running main.py directly):
# OLLAMA_BASE_URL=http://localhost:11434
# DEFAULT_MODEL=ollama/llama3.2
#
# For remote Ollama server:
# OLLAMA_BASE_URL=http://your-ollama-server:11434
# DEFAULT_MODEL=ollama/llama3.1:70b
#
# Popular Ollama models:
# ollama/llama3.2 # 3B params, fast
# ollama/llama3.1:8b # 8B params, good balance
# ollama/llama3.1:70b # 70B params, high quality
# ollama/mistral # 7B params, fast
# ollama/mixtral # 8x7B MoE, powerful
# ollama/deepseek-r1:14b # DeepSeek reasoning model
# ollama/qwen2.5:72b # Qwen 2.5, excellent for structured output
#
# OLLAMA_BASE_URL=http://host.docker.internal:11434
# DEFAULT_MODEL=ollama/llama3.2