Spaces:
Configuration error
Configuration error
File size: 3,491 Bytes
762ef74 24923c7 762ef74 24923c7 762ef74 |
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 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 |
# full.llama.cpp.agent.python
# 🦙 LLaMA Local AI Agent
A **fully offline**, powerful, and extendable AI agent that runs **LLaMA 3/gguf models locally** via `llama.cpp`. Supports **Chat, Document Q&A, Vision (OCR + Image Captioning), Voice (STT & TTS)**, and **Tools** like Python exec, browser search, and email generation — all from your machine.
---
## 🧠 Features
- 🔁 LLaMA-3 / GGUF Models (Runs Locally via `llama.cpp`)
- 💬 Chat Interface with History & Token Auth
- 📄 File Upload + Document Q&A (TXT/PDF)
- 🧰 Tools: Calculator, Python Exec, Joke, Web Search
- 🖼️ Image Captioning + OCR
- 🗣️ Voice Transcription (Whisper) + Text-to-Speech (TTS)
- 🔐 JWT Auth for Admin/Users
- 📜 Chat Export with Embedding Search
- 📊 Agent Analytics & Logs
- 🧭 ReAct / LangGraph Planning (Coming Soon)
---
## 🚀 Quick Start
```bash
git clone https://github.com/yourusername/llama-agent
cd llama-agent
docker-compose up --build
```
Then go to `http://localhost:8501` to access the Streamlit UI.
---
## 🧩 Architecture
- `main_api.py`: FastAPI backend (chat, tools, auth, vector store)
- `streamlit_app.py`: Rich UI frontend
- `llama.cpp`: LLM inference backend via `llama_cpp` Python bindings
- `whisper`, `pyttsx3`: Voice STT/TTS
- `duckduckgo_search`, `playwright`: Web tools
- `faiss`, `sentence-transformers`: Vector search
- `PDFMiner`, `Pillow`, `Tesseract`: PDF/Image Processing
---
## 📸 Screenshots
<p align="center">
<img src="screenshots/agent-ui.png" width="600" />
<img src="screenshots/chat-demo.png" width="600" />
</p>
---
## 🔧 Tools
### Tool Dispatch Logic
```python
from app.repl_tool import run_code
from app.email_tool import send_email
from app.translation_tool import translate
from app.vision import describe_image
def tool_dispatch(command, args):
if command == "run_code":
return run_code(args)
if command == "send_email":
return send_email(*args)
if command == "translate":
return translate(args)
if command == "describe_image":
return describe_image(args)
return "❌ Unknown command"
```
---
## 🛠️ Logging
```python
import logging
logging.basicConfig(
filename="agent.log",
level=logging.INFO,
format="%(asctime)s | %(levelname)s | %(message)s",
)
def log_event(event: str):
logging.info(event)
```
---
## 📤 Email Tool
```python
from email.message import EmailMessage
import smtplib, ssl
EMAIL_SENDER = "[email protected]"
EMAIL_PASSWORD = "yourpassword"
def generate_email(recipient_name: str, product_name: str, discount: float):
return f"""
Dear {recipient_name},
Great news! We're excited to offer you a special {discount}% discount on our premium product: {product_name}.
Visit our site to take advantage of this limited-time offer.
Best regards,
Your AI Assistant
"""
def send_email(to, subject, body):
msg = EmailMessage()
msg.set_content(body)
msg["Subject"] = subject
msg["From"] = EMAIL_SENDER
msg["To"] = to
with smtplib.SMTP_SSL("smtp.gmail.com", 465, context=ssl.create_default_context()) as smtp:
smtp.login(EMAIL_SENDER, EMAIL_PASSWORD)
smtp.send_message(msg)
return "✅ Email sent"
```
---
## 📝 License
MIT License — Open-source, feel free to extend it.
---
## 🙏 Credits
- [`llama.cpp`](https://github.com/ggerganov/llama.cpp)
- [`whisper`](https://github.com/openai/whisper)
- [`Streamlit`](https://streamlit.io) |