Reki Bets

Reki Bets is a two-agent system for sports betting analysis and supporting research. The repo contains:

• Agent 1 (Betting Analyst): A FastAPI service that uses sports data tools (NFL/NBA) and a Streamlit UI to deliver betting analysis.
• Agent 2 (Research Agent): A FastAPI service that uses Brave Search plus Grok for web research and summarization, with a scheduler to trigger daily runs.

Repository layout

.
├── agent1/                 # Betting analyst API + Streamlit UI
├── agent2/                 # Research agent API + scheduler
└── spot.txt                # Project status and current prioritie

Prerequisites

• Python 3.12+ for agent1 (per agent1/pyproject.toml).
• Python 3.9+ for agent2 (per agent2/pyproject.toml).
• API keys for the external services listed below.

Environment variables

Create .env files in each agent directory (or export these variables in your shell). The services will error at startup if required values are missing.

Agent 1

Required:

• GEMINI_API_KEY
• BRAVE_API_KEY
• SPORTRADAR_API_KEY
• SERPAPI_API_KEY

Optional:

• TIMEZONE (defaults to UTC)
• XAI_API_KEY (used if calling Grok in the agent)
• XAI_BASE_URL (used if calling Grok in the agent)

Agent 2

Required:

• BRAVE_API_KEY
• XAI_API_KEY
• XAI_BASE_URL

Running Agent 1 (Betting Analyst)

From agent1/, install dependencies and start the FastAPI service:

cd agent1
python -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install -r <(python - <<'PY'
import tomllib
from pathlib import Path
pyproject = tomllib.loads(Path('pyproject.toml').read_text())
print("\n".join(pyproject["project"]["dependencies"]))
PY
)

python api.py

In a separate terminal, run the Streamlit UI:

cd agent1
source .venv/bin/activate
streamlit run ui.py

The API serves OpenAI-compatible streaming responses at http://127.0.0.1:8005/v1/chat/completions, which the Streamlit UI consumes.

Running Agent 2 (Research Agent)

From agent2/, install dependencies with Poetry and start the FastAPI service:

cd agent2
poetry install
poetry run uvicorn researcher:app --host 0.0.0.0 --port 8007

To trigger the scheduled daily research job:

cd agent2
poetry run python scheduler.py

Project status

See spot.txt for current priorities including odds tooling fixes for Agent 1 and stabilization work for Agent 2’s scraping/summarization pipeline.