fc5335a5c6
- 将 test_ui/__init__.py 中的测试代码移至 test_ui/test_ui.py - 将 test_integration/__init__.py 中的测试代码移至 test_integration/test_integration.py - 删除 test_packet.py 中重复的 TestToolLimiter 和 TestEmbeddedGraphDB - 将 TestPacketTypeEnum 的 9 个重复测试合并为参数化测试 - 移除 conftest.py 中从未使用的 6 个 fixtures - 修复 3 个预存测试 bug (update_config 不存在、limiter 断言、error 处理)
TrulyMEM - TrueHumanMEM
📜 License: GNU General Public License v3.0 (GPLv3)
This project is free and open source. You are free to use, modify, and distribute, but modified works must be distributed under the same license.
中文: 切换到中文版
Give AI self-awareness, plasticity, and a sense of proportion in long-term memory
The More Human Choice.
⚠️ Current Branch:
test— Testing branch for experimental features. Code here may be unstable and should not be used in production. For stable releases, see themainbranch.
The Story
Industry believes that LLMs' massive parameters give them emergent intelligence. But this intelligence is "dead" — it cannot truly remember, nor understand the concept of "remembering". Everything it outputs is the probabilistic optimal solution calculated through countless forward passes on the current input text. The LLM cannot correct its weights based on errors in a conversation, nor perform a backward pass. Its consciousness is frozen — what appears as intelligence is merely the echo of this frozen consciousness.
Current "memory systems" merely externalize memory, letting the "system" remember for the LLM. Or they dump all context text to the LLM. This is a waste of the model's limited input context.
TrulyMEM asks: since the LLM cannot correct model weights in real-time, why not give the memory authority back to the LLM?
We provide a series of mechanisms for the LLM to decide what to remember, what to forget, what's important, what's trivial. The LLM's reasoning process is also its thinking and recalling process. Abandoning the traditional messages array context, all memories are stored as triplets (graph) in the graph database. When the LLM thinks, it can autonomously jump through graph links to associate related relationships, enabling natural association and recall.
Give the LLM true memory.
Quick Start
Method 1: Run Packaged Executable
# Windows: TrulyMEM.exe
# Linux/macOS: TrulyMEM
chmod +x TrulyMEM
./TrulyMEM
Method 2: Run from Source
git clone <repo-url>
cd TrulyMEM-TrueHumanMEM
pip install -r requirements.txt
python trulymem_entry.py
Configuration
- Press F2 to expand sidebar
- Enter API Key (supports DeepSeek, OpenAI, etc.)
- Press Enter to save
- Start chatting!
Documentation Index
Detailed technical documentation in the docs/en/ directory:
| Document | Content |
|---|---|
| docs/en/architecture.md | System architecture and technical design |
| docs/en/quick_start.md | Complete startup guide and configuration |
| docs/en/memory.md | Internal memory working mechanism |
| docs/en/persona.md | Persona Graph mechanism |
| docs/en/working_memory.md | Continuous task handling mechanism |
| docs/en/api.md | BackendServer API (for extension development) |
| docs/en/prompts.md | Prompt management module |
Contributing
Welcome to submit Issues and Pull Requests!
- Fork this repository
- Create feature branch (
git checkout -b feature/AmazingFeature) - Commit changes (
git commit -m 'Add some AmazingFeature') - Push to branch (
git push origin feature/AmazingFeature) - Create Pull Request
License
This project is licensed under the GNU General Public License v3.0 (GPLv3).
See LICENSE file for details.
Special Thanks
- Prof. Meiting Wang — Academic guidance
- 逝水秋生白 — Architecture support
- anzhitinglan — Testing resource support
- 崔莉萍老师 — Theoretical guidance
- Annie — Professional guidance
- 王梓沣、马悦华、隆梦婷 — Neuroscience theory support