AgentX-Python/.env.example at main · AgentX-ai/AgentX-Python · GitHub

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
# AgentX
AGENTX_API_KEY=your_agentx_api_key_here
# Optional: override the API base URL (defaults to production)
# AGENTX_API_BASE_URL=https://api.agentx.so/api/v1/

# ── LLM providers ───────────────────────────────────────────────────────────

# OpenAI — used by: openai_eval.py, langchain_eval.py, llamaindex_eval.py,
#                   crewai_eval.py, autogen_eval.py, multi_model_eval.py
OPENAI_API_KEY=sk-...

# Anthropic — used by: anthropic_eval.py, multi_model_eval.py
ANTHROPIC_API_KEY=sk-ant-...

# Google AI (Gemini) — used by: multi_model_eval.py
# Note: install "google-genai" (not the deprecated "google-generativeai")
GOOGLE_API_KEY=AIza...

# ── Per-example mode toggles ─────────────────────────────────────────────────
# Uncomment and set before running the relevant script.

# openai_eval.py       — "chat" (gpt-4o-mini + tools) | "reason" (o4-mini)
# OPENAI_MODE=chat

# anthropic_eval.py    — "standard" (haiku) | "thinking" (opus-4-7)
# ANTHROPIC_MODE=standard

# langchain_eval.py    — "chain" (LCEL) | "agent" (ReAct + tools)
# LANGCHAIN_MODE=chain

# llamaindex_eval.py   — "rag" (RAG pipeline) | "agent" (ReAct agent)
# LLAMA_MODE=rag

# crewai_eval.py       — "single" | "multi" (researcher + writer crew)
# CREWAI_MODE=single

# autogen_eval.py      — "single" | "group" (GroupChat)
# AUTOGEN_MODE=single

# multi_model_eval.py  — comma-separated subset, e.g. "openai,anthropic"
#                        leave unset to run all providers
# PROVIDERS=openai,anthropic,google