https://github.com/towry/fund-trend
Fund trend predication with llm agent, Toy with llm agent, do not take it serious :)
https://github.com/towry/fund-trend
funds goose llm-agents sub-agents
Last synced: 2 months ago
JSON representation
Fund trend predication with llm agent, Toy with llm agent, do not take it serious :)
- Host: GitHub
- URL: https://github.com/towry/fund-trend
- Owner: towry
- Created: 2025-07-31T16:51:12.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-08-12T00:57:53.000Z (10 months ago)
- Last Synced: 2025-10-13T20:22:37.021Z (8 months ago)
- Topics: funds, goose, llm-agents, sub-agents
- Language: Shell
- Homepage:
- Size: 76.2 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 11
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGES.md
Awesome Lists containing this project
README
# Fund Trend Analysis
This repository contains a GitHub Actions workflow that automatically predicts the movement of the Nasdaq-100 index to guide investment decisions.
> It schedule on every day at 9 PM UTC
## Purpose
The goal of this project is to:
- Automatically collect the latest financial news related to the Nasdaq-100 index
- Analyze the news using AI with financial expertise
- Predict the short-term direction of the Nasdaq-100 index
- Provide data-driven guidance for investment decisions
## How It Works
1. The workflow runs daily (or can be triggered manually)
2. It uses the Perplexity search tool to gather the latest relevant news
3. The collected news is analyzed using a financial AI model
4. A prediction is generated with supporting rationale
5. Results are saved and stored in this repository
## Directory Structure
- `news/`: Stores collected news articles
- `predictions/`: Stores prediction results
- `.github/workflows/`: Contains the GitHub Actions workflow
- `scripts/`: Contains helper scripts
## Getting Started
To use this repository, you need to:
1. Set up an OpenRouter API key from [OpenRouter](https://openrouter.ai/)
2. Set up a Perplexity API key from [Perplexity AI](https://www.perplexity.ai/)
3. Add them as GitHub secrets named `OPENROUTER_API_KEY` and `PERPLEXITY_API_KEY`
4. The workflow will automatically run based on the schedule
## Local Development
To run the prediction script locally:
1. Install dependencies:
```bash
pip install -r requirements.txt
```
2. Set environment variables:
```bash
export OPENROUTER_API_KEY=your_openrouter_api_key
export PERPLEXITY_API_KEY=your_perplexity_api_key
```
3. Run the prediction script:
```bash
./scripts/predict-nasdaq.sh
```
## Configuration
### Required GitHub Secrets
Before running the workflow, you need to configure the following secrets in your GitHub repository:
| Secret Name | Description | How to Obtain |
| -------------------- | ---------------------------------------- | ---------------------------------------------------------------------------- |
| `OPENROUTER_API_KEY` | API key for OpenRouter AI service | Get from [OpenRouter dashboard](https://openrouter.ai/keys) |
| `PERPLEXITY_API_KEY` | API key for Perplexity AI search | Get from [Perplexity API](https://docs.perplexity.ai/guides/getting-started) |
| `PUSHOVER_USER` | Your Pushover user key for notifications | Get from [Pushover dashboard](https://pushover.net/) |
| `PUSHOVER_TOKEN` | Your Pushover application token | Create at [Pushover apps](https://pushover.net/apps) |
### How to Set GitHub Secrets
1. Go to your GitHub repository
2. Navigate to **Settings** → **Secrets and variables** → **Actions**
3. Click **New repository secret**
4. Add each secret with its corresponding value
### Workflow Configuration
The workflow uses Goose CLI with OpenRouter as the provider and Perplexity MCP server for searching. The configuration includes:
- **OpenRouter Provider**: For accessing various AI models
- **Perplexity MCP Server**: For gathering the latest financial news and market analysis
- **Goose Models**: Configurable AI models for general processing and financial analysis
- **Scheduled Runs**: Weekly execution on Fridays at 9 AM UTC with optional manual triggering
## Customization
You can modify:
- The schedule in `.github/workflows/predict-nasdaq.yml`
- The search query used to find news
- The AI model used for analysis
- The prediction criteria
- Pushover notification settings in the workflow file
- The search query used to find news
- The AI model used for analysis
- The prediction criteria
## License
MIT License