Spaces:
Running
Running
intro & instructions
Browse files- README.md +45 -1
- readme-assets/claude-slide-translation.webp +0 -0
README.md
CHANGED
|
@@ -12,7 +12,51 @@ tags:
|
|
| 12 |
- building-mcp-track-customer
|
| 13 |
---
|
| 14 |
|
| 15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
## Local development
|
| 18 |
|
|
|
|
| 12 |
- building-mcp-track-customer
|
| 13 |
---
|
| 14 |
|
| 15 |
+
# Aileen 3 Core
|
| 16 |
+
## Introduction
|
| 17 |
+
Large Language Models (LLMs) rely on **tools** - sometimes provided by **MCP servers** - to interact with the outside world. Aileen 3 Core is an MCP server that advances the goals of **Information Foraging**: mining for novel insights from high-noise sources to create dense briefings for time efficient consumption by the user.
|
| 18 |
+
|
| 19 |
+
Aileen 3 Core is [a contender](https://huggingface.co/spaces/ndurner/aileen3-core) in the [MCP's 1st Birthday - Hosted by Anthropic and Gradio](https://huggingface.co/MCP-1st-Birthday) hackathon. [Aileen 3 Agent](https://github.com/ndurner/aileen3-agent), an agentic system built on this MCP server, is a [capstone project](https://ndurner.de/links/aileen3-kaggle-writeup) to the [AI Agents Intensive Course with Google](https://www.kaggle.com/learn-guide/5-day-agents).
|
| 20 |
+
|
| 21 |
+
## Using
|
| 22 |
+
### Using in Claude Desktop
|
| 23 |
+
#### Installing
|
| 24 |
+
1. Optionally, create a new Python virtual environment
|
| 25 |
+
* example: `python3 -m venv .venv-claude`
|
| 26 |
+
2. Install the Aileen MCP: `pip install ./mcp`
|
| 27 |
+
* (or `pip install -e ./mcp` if you want to make live changes to this source tree)
|
| 28 |
+
3. Obtain Google Gemini API key: [https://aistudio.google.com](Google AI Studio)
|
| 29 |
+
3. Add reference to `claude_desktop_config.json`. The Gemini API key will be read from the environment, so can be set here:
|
| 30 |
+
```
|
| 31 |
+
{
|
| 32 |
+
...
|
| 33 |
+
"mcpServers": {
|
| 34 |
+
"aileen3-mcp": {
|
| 35 |
+
"command": "/Users/.../aileen3-core/.venv-claude/bin/python",
|
| 36 |
+
"args": [
|
| 37 |
+
"-m",
|
| 38 |
+
"aileen3_mcp.server"
|
| 39 |
+
],
|
| 40 |
+
"env": {
|
| 41 |
+
"GEMINI_API_KEY": "AI..."
|
| 42 |
+
}
|
| 43 |
+
}
|
| 44 |
+
}
|
| 45 |
+
}
|
| 46 |
+
```
|
| 47 |
+
4. Restart Claude
|
| 48 |
+
|
| 49 |
+
#### Using the MCP server
|
| 50 |
+
The model Haiku 4.5 is sufficient for basic tasks. To make your plans fully transparent to the LLM, refer to "aileen3" explicitely in the prompt, e.g.:
|
| 51 |
+
> Use aileen3 to translate slide 3 from YouTube video reference eXP-PvKcI9A to German.
|
| 52 |
+
|
| 53 |
+

|
| 54 |
+
|
| 55 |
+
#### Debugging
|
| 56 |
+
The message exchange and Claude-facing error messages can be read from Claude log files:
|
| 57 |
+
```
|
| 58 |
+
tail -n 20 -F ~/Library/Logs/Claude/mcp*.log
|
| 59 |
+
```
|
| 60 |
|
| 61 |
## Local development
|
| 62 |
|
readme-assets/claude-slide-translation.webp
ADDED
|