https://github.com/deepleaper/leaper-agent
Self-learning AI agent framework with 6-layer memory evolution engine. Build intelligent agents for any business scenario.
https://github.com/deepleaper/leaper-agent
ai-agent ai-employee cxo memory python self-evolving telegram
Last synced: 16 days ago
JSON representation
Self-learning AI agent framework with 6-layer memory evolution engine. Build intelligent agents for any business scenario.
- Host: GitHub
- URL: https://github.com/deepleaper/leaper-agent
- Owner: Deepleaper
- License: other
- Created: 2026-04-26T05:08:11.000Z (3 months ago)
- Default Branch: master
- Last Pushed: 2026-05-13T10:39:24.000Z (2 months ago)
- Last Synced: 2026-05-13T12:35:11.038Z (2 months ago)
- Topics: ai-agent, ai-employee, cxo, memory, python, self-evolving, telegram
- Language: Python
- Homepage: https://github.com/Deepleaper
- Size: 13.5 MB
- Stars: 2
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Security: SECURITY.md
- Agents: AGENTS.md
Awesome Lists containing this project
README
# ๐ Leaper Agent
### The Agent Framework That Actually Remembers
### Hermes-Compatible Agent Orchestration + DeepBrain 6-Layer Evolving Memory
[](https://pypi.org/project/leaper-agent/)
[](https://pypi.org/project/leaper-agent/)
[](https://github.com/deepleaper/leaper-agent/stargazers)
[](LICENSE)
[](https://python.org)
[Website](https://www.deepleaper.com) ยท [Quick Start](#-quick-start) ยท [Why Memory Matters](#-why-memory-matters) ยท [ไธญๅฝ็](https://github.com/deepleaper/leaper-agent-cn)
---
## ๐ค The Problem
Every agent framework gives you tool-calling and multi-model support. **None of them give you real memory.**
- CrewAI? Flat key-value store, resets between runs.
- AutoGen? Conversation history only, no knowledge evolution.
- LangGraph? Checkpoints, not understanding.
- Hermes Agent? Skill files on disk โ smart, but static.
**Your agent forgets everything the moment the session ends.**
## ๐ก The Solution
**Leaper Agent = [Hermes Agent](https://github.com/hermes-agent) base + [DeepBrain](https://github.com/deepleaper/opc-deepbrain) memory engine.**
We took the proven Hermes Agent architecture โ its agent loop, tool runtime, and multi-provider support โ and integrated DeepBrain's 6-layer self-evolving memory engine. The result: **an agent that gets smarter every time you talk to it.**
```
Session 1: "I'm building a SaaS product with FastAPI and React."
Session 2: "What stack should I use for the admin panel?"
โ Agent recalls your tech choices, suggests FastAPI Admin + React
โ No manual context injection. It just knows.
```
## ๐ง Why Memory Matters
| What happens | Without memory | With DeepBrain |
|-------------|---------------|----------------|
| You mention your tech stack | Forgotten next session | Remembered forever, evolves over time |
| You correct the agent | Same mistake tomorrow | Learns the correction, never repeats |
| You have 50 conversations | Each starts from zero | Agent builds a knowledge graph of YOU |
| Knowledge conflicts | Silently contradicts itself | 4-Gate system detects & resolves conflicts |
### 6-Layer Memory Architecture
```
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ Layer 5: Meta-Knowledge โ
โ "I know your tech preferences well, โ
โ but I'm uncertain about your budget" โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Layer 4: Archived โ Historical ref โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Layer 3: Consolidated โ Cross-session โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Layer 2: Long-Term โ Validated facts โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Layer 1: Short-Term โ Recent context โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ Layer 0: Flash โ Current session โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ Auto-promotion via 4-Gate QC โ
(Relevance ยท Novelty ยท Consistency ยท Utility)
```
This isn't RAG. This isn't vector search. This is **knowledge that evolves** โ facts get validated, promoted, consolidated, and the agent develops meta-awareness of what it knows well and what it doesn't.
## ๐ Quick Start
```bash
# Install
pip install leaper-agent
# Set your API key
export OPENAI_API_KEY=sk-xxx
# or: export ANTHROPIC_API_KEY=sk-ant-xxx
# or: export GEMINI_API_KEY=xxx
# Create an agent with the CEO Coach template
leaper create my-agent --template ceo-coach
# Start
leaper start my-agent
```
### Deploy as Telegram Bot
```bash
leaper create my-bot --template ceo-coach --bot-token YOUR_TELEGRAM_TOKEN
leaper start my-bot
```
## ๐๏ธ Architecture
```
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ Leaper Agent Runtime โ
โ (Hermes-compatible agent loop) โ
โโโโโโโโโโโฌโโโโโโโโโโโฌโโโโโโโโโโโฌโโโโโโโโโค
โ OpenAI โ Claude โ Gemini โ Ollama โ
โโโโโโโโโโโดโโโโโโโโโโโดโโโโโโโโโโโดโโโโโโโโโค
โ Agent Orchestration โ
โ Role templates ยท Tool calling โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ ๐ง DeepBrain Memory Engine โ
โ 6-Layer ยท 4-Gate ยท Zero Dependencies โ
โ SQLite-only ยท 100% Local ยท Evolving โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
```
**Key difference from Hermes:** Hermes stores skills as static Markdown files. Leaper Agent's DeepBrain **automatically extracts, validates, and evolves knowledge** from every conversation โ no manual curation needed.
## โก Hermes Compatibility
Leaper Agent is built on the Hermes Agent foundation. If you're familiar with Hermes, you'll feel at home:
- Same agent loop architecture
- Same multi-provider model support
- Same tool runtime
- **Plus:** 6-layer evolving memory that Hermes doesn't have
## ๐ Supported Providers
| Provider | Models | Status |
|----------|--------|--------|
| OpenAI | GPT-4o, GPT-4o-mini, o1/o3 | โ
|
| Anthropic | Claude Sonnet, Opus, Haiku | โ
|
| Google | Gemini 2.0, 2.5 | โ
|
| Ollama | Any local model | โ
|
## ๐ vs Other Frameworks
| | Leaper Agent | Hermes | CrewAI | AutoGen | LangGraph |
|---|:---:|:---:|:---:|:---:|:---:|
| Memory that evolves | โ
6-layer | โ Static files | โ Flat KV | โ Chat history | โ Checkpoints |
| Knowledge quality control | โ
4-Gate | โ | โ | โ | โ |
| Meta-knowledge | โ
| โ | โ | โ | โ |
| Multi-provider | โ
| โ
| โ
| โ
| โ
|
| Zero memory dependencies | โ
SQLite only | N/A | โ Redis/Qdrant | โ | โ Vector DB |
| Hermes compatible | โ
| โ
| โ | โ | โ |
## ๐จ๐ณ China Version / ไธญๅฝ็
For Chinese LLM providers (DeepSeek, ้ไนๅ้ฎ, ๆบ่ฐฑ, Moonshot) + 150 role templates:
```bash
pip install leaper-agent-cn
```
โ [Leaper Agent CN](https://github.com/deepleaper/leaper-agent-cn)
## ๐ Ecosystem
| Project | Description |
|---------|-------------|
| [OPC DeepBrain](https://github.com/deepleaper/opc-deepbrain) | Standalone memory engine (use in any framework) |
| [OPC Agent](https://github.com/deepleaper/opc-agent) | Local-first agent with Ollama |
| [Leaper Agent CN](https://github.com/deepleaper/leaper-agent-cn) | China-optimized with 150 templates |
## ๐ License
[BSL-1.1](LICENSE) โ Free for non-competitive use. Converts to Apache-2.0 after 4 years.
## ๐ค Contributing
We welcome contributions! See [CONTRIBUTING.md](CONTRIBUTING.md).
๐ง Contact: [tech@deepleaper.com](mailto:tech@deepleaper.com)
---
**Built with โค๏ธ by [Deepleaper Technology / ่ท็็งๆ](https://www.deepleaper.com)**
*Other agents forget. Leaper Agent evolves.*