| name | LatteReview (Systematic Review Automation) |
| description | Automates systematic literature reviews using multi-agent AI workflows.
Can perform title/abstract screening, full-text screening, and data abstraction.
|
| author | BioDockify AI Team (Integration of PouriaRouzrokh/LatteReview) |
| version | 1.0.0 |
LatteReview Skill
This skill integrates the LatteReview framework to allow BioDockify AI to perform PhD-level systematic reviews.
Capabilities
- Screening: Filter papers based on inclusion/exclusion criteria.
- Scoring: Verify relevance with confidence scores.
- Abstraction: Extract structured data points from papers.
Usage in BioDockify AI
from agent_zero.skills.latte_review import get_latte_review
latte = get_latte_review()
inclusion = "Studies using AI for drug discovery in oncology."
exclusion = "Review articles, non-English papers."
output_csv = latte.screen_papers(
input_path="data/papers/candidates.csv",
inclusion_criteria=inclusion,
exclusion_criteria=exclusion
)
print(f"Screening complete. Results saved to {output_csv}")
Configuration
The skill uses LiteLLM under the hood. It will inherit the API keys from the BioDockify AI environment (e.g., OPENAI_API_KEY, GEMINI_API_KEY).
