MemoryAgent/.env.example at main · sandy1990418/MemoryAgent · 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
# Required for demos that call OpenAI-compatible LangChain models.
OPENAI_API_KEY=

# Shared model knobs.
MAIN_MODEL=openai:gpt-5.4-nano
SUMMARY_MODEL=openai:gpt-5.4-nano
MEMORY_MODEL=openai:gpt-5.4-nano
THREAD_ID=react-summary-demo

# Structured-memory prompt/window settings.
MAX_WINDOW_TOKENS=300
STRUCTURED_MAX_TOKENS=600
STRUCTURED_MAX_MEMORY_TOKENS=600
STRUCTURED_KEEP_MESSAGES=4
# practical (default), agent, eval, or beam (alias for eval).
MEMORY_PROFILE=practical
MEMORY_SECTIONS=practical
MEMORY_COMPACTION_THRESHOLD=30

# mem0 backend:
# - local: development/test mode. Uses MEM0_DATA_DIR and embedded Qdrant.
# - platform: your hosted/custom mem0 data. MEM0_DATA_DIR is ignored.
# - disabled: structured memory only; no long-term vector memory.
MEM0_BACKEND=local
MEM0_USER_ID=demo-user
MEM0_DATA_DIR=.mem0
MEM0_LLM_MODEL=gpt-5.4-nano

# Required only when MEM0_BACKEND=platform.
MEM0_API_KEY=

# BEAM runner overrides.
BEAM_CONFIG=configs/beam.yaml
BEAM_DATA_PATH=BEAM/chats/100K/1
BEAM_ABILITIES=contradiction_resolution,knowledge_update,preference_following,instruction_following,abstention,summarization
BEAM_JUDGE=true
BEAM_MAX_QUESTIONS_PER_TYPE=
BEAM_ANSWER_MODEL=gpt-5.4-nano
BEAM_MEMORY_MODEL=gpt-5.4-nano
BEAM_JUDGE_MODEL=