Maxim生物启发认知架构 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 7.2 分,整体质量较高。我们推荐使用将其纳入你的 AI 工具库,帮助提升工作效率。
Maxim生物启发认知架构 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
Maxim生物启发认知架构 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install maxim
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install maxim
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/dennys246/Maxim
cd Maxim
pip install -e .
# 验证安装
python -c "import maxim; print('安装成功')"
# 命令行使用
maxim --help
# 基本用法
maxim input_file -o output_file
# Python 代码中调用
import maxim
# 示例
result = maxim.process("input")
print(result)
# maxim 配置文件示例(config.yml) app: name: "maxim" debug: false log_level: "INFO" # 运行时指定配置文件 maxim --config config.yml # 或通过环境变量配置 export MAXIM_API_KEY="your-key" export MAXIM_OUTPUT_DIR="./output"
Bio-inspired cognitive harness for LLM agents — embodied sensation, homeostatic drives, and brain-modeled persistent memory let LLM-driven agents carry learning across sessions without fine-tuning.
Maxim gives an LLM agent a body (sensors, modulators, pain), drives (hunger, temperature, fatigue that drift and compete), and biological memory systems (Hippocampus, NAc, ATL, SCN, Angular Gyrus) that capture experience. When the agent's body touches fire, its thermal sensors register pain, NAc forms a causal link, and the enrichment pipeline surfaces that experience in subsequent sessions — providing the LLM with experience-grounded context alongside its pretraining. The bio-substrate doesn't replace the LLM's prior knowledge; it augments the LLM's prompt context with persistent, agent-specific lived experience.
Positioning (per Exp 37 2026-06-06 results): Maxim is a bio-inspired LLM harness. The substrate provides cross-session infrastructure (memory, valence, causal links, drives) that LLM-driven agents use. Substrate-driven action selection independent of the LLM is post-1.0 research direction via Exp 38 substrate-primary work. See docs/plans/behavioral_graduation_candidates.md for the Tier 1 graduation status.
Website: dennyschaedig.com/maxim
hunger: drive: drift_mode: entropic drift_direction: up drift_rate: 0.006 deprivation_threshold: 0.7 deprivation_pain: 0.3 ```
Three sensation layers converge on the same pipeline: - Contact (entity acquisition): pick up a rock → its sensors join your body → damage model evaluates - Touch (self_effect): touch fire → one-time thermal spike on arms - Narrative (keyword reflexes): narrator describes flames → reflex fires → damage → pain
All produce: sensor change → evaluate_failures() → PainBus → NAc learning.
maxim.run(model="mistral-7b")
pip install 'pymaxim[llm-llama,llm-server]' maxim --list-models # see available models maxim --sim "test memory recall" --llm mistral-7b # auto-downloads on first run
maxim --sim cradle --embodiment bodies/infant_humanoid --sim-max-turns 25 ```
Check your setup with maxim doctor, and find simulation reports in ~/.maxim/sim_reports/{session_id}/.
pip install pymaxim
import maxim
```bash
The dennyschaedig.com/maxim site hosts long-form topic guides covering the full architecture. These are the deep-dive companion to this README:
| Guide | Topic |
|---|---|
| [Overview](https://dennyschaedig.com/maxim/maxim-overview) | Project introduction and design philosophy |
| [Agent Architecture](https://dennyschaedig.com/maxim/maxim-agent-architecture) | Layered architecture, bio-system pipeline, fear circuit, cerebellum |
| [Memory Systems](https://dennyschaedig.com/maxim/maxim-memory-systems) | Hippocampus, NAc, SCN, ATL, EC, Angular Gyrus — in depth |
| [Semantic Memory](https://dennyschaedig.com/maxim/maxim-semantic-memory) | ATL concept formation and retrieval |
| [Embodiment](https://dennyschaedig.com/maxim/maxim-embodiment) | Sensor-Entity-Modulator protocol, drives, pain cascade |
| [Proprioception & Body Awareness](https://dennyschaedig.com/maxim/maxim-proprioception) | Body state, drive evaluation, interoception |
| [Prompt System & Tool Injection](https://dennyschaedig.com/maxim/maxim-prompt-system) | How memory and substrate enrich LLM context |
| [Deliberation](https://dennyschaedig.com/maxim/maxim-deliberation) | PFC inner monologue and thought stream |
| [Imagination](https://dennyschaedig.com/maxim/maxim-imagination) | Real-time entity design from novel percepts |
| [Simulation](https://dennyschaedig.com/maxim/maxim-simulation) | Percept simulation, scenario testing |
| [DM Campaigns](https://dennyschaedig.com/maxim/maxim-dm-campaigns) | Multi-encounter branching stories |
| [Benchmarks](https://dennyschaedig.com/maxim/maxim-benchmarks) | Multi-model cognitive testing |
| [Component Library](https://dennyschaedig.com/maxim/maxim-component-library) | SEM entity templates and genres |
| [Concept Decomposition](https://dennyschaedig.com/maxim/maxim-concept-decomposition) | Noun-phrase extraction for substrate encoding |
| [Attention & Salience](https://dennyschaedig.com/maxim/maxim-attention-salience) | Salience modulation and attention weighting |
| [Operating Modes](https://dennyschaedig.com/maxim/maxim-operating-modes) | Autonomy levels, sleep, planning vs supervised |
| [Networking](https://dennyschaedig.com/maxim/maxim-networking) | Multi-LLM distributed inference, leader/peer setup |
| [Agent Mesh](https://dennyschaedig.com/maxim/maxim-agent-mesh) | Peer-to-peer knowledge sharing (Hivemind) |
| [Hivemind + Oasis](https://dennyschaedig.com/maxim/maxim-hivemind) | Federated bio-substrate architecture |
| [Substrate-Primary Mode](https://dennyschaedig.com/maxim/maxim-substrate-primary) | Bio-substrate driven action selection |
| [Tools & Introspection](https://dennyschaedig.com/maxim/maxim-tools) | Agent tool system and discovery |
| [Math & Statistical Cognition](https://dennyschaedig.com/maxim/maxim-math-cognition) | Statistician agent, variance, NAc reward |
| [Experiments & Results](https://dennyschaedig.com/maxim/maxim-experiments) | Bio-inspired AI research findings |
| [Technical Deep Dive](https://dennyschaedig.com/maxim/maxim-technical-deepdive) | Architecture, threading, persistence |
| [Usage Guide](https://dennyschaedig.com/maxim/maxim-usage-guide) | Install, config, and CLI walkthrough |
| [Roadmap](https://dennyschaedig.com/maxim/maxim-roadmap) | Future plans and development milestones |
| Extra | What it adds |
|---|---|
llm-llama | Local LLM inference via llama.cpp |
llm-torch | PyTorch/Transformers backend |
llm-anthropic | Claude backend |
llm-openai | OpenAI backend |
vision | Camera + object detection |
audio | Microphone + Whisper transcription |
reachy | Reachy Mini robot SDK |
comms | Twilio SMS/Voice |
semantic | Sentence-transformer embeddings |
tts | Text-to-speech via Piper |
database | PostgreSQL + pgvector memory stores |
See getting-started.md for the full list of 16 extras.
Note:[all]does not include[semantic](sentence-transformers + spaCy). Without it, memory recall and substrate encoding fall back to bag-of-words hashing. For full memory quality:> pip install 'pymaxim[all,semantic]' >
```bash
report = maxim.diagnose()
maxim config list # show all resolved settings maxim config get lanes.large.remote_url # get a single field maxim config set cloud.enabled true # set a field
```python
```bash
创新的生物启发认知框架,将embodied AI理念融入LLM工作流。架构设计前沿但成熟度有限,适合学术研究和前沿探索。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ Apache 2.0 — 宽松开源协议,可商用,需保留版权声明和 NOTICE 文件,含专利授权条款。
经综合评估,Maxim生物启发认知架构 在Agent工作流赛道中表现稳健,质量良好。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | Maxim |
| 原始描述 | 开源AI工作流:Bio-inspired cognitive architecture for LLM agents providing embodied sensation,。⭐6 · Python |
| Topics | 生物启发认知架构LLM智能体具身AI记忆系统 |
| GitHub | https://github.com/dennys246/Maxim |
| License | Apache-2.0 |
| 语言 | Python |
收录时间:2026-06-04 · 更新时间:2026-06-11 · License:Apache-2.0 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端