Automate your systematic literature reviews with AI — from import to extraction table.
🚀 Quick Start · 🌐 Web App · 📖 Full Guide · 🐛 Issues · 💡 Features
A powerful, open-source tool for automating systematic literature reviews (SLRs) with support for 9 AI providers, reference ingestion, deduplication, two-stage screening, and structured data extraction. Available as both a desktop GUI and a browser-based Web App.
New to this tool? Read COMPLETE_USER_GUIDE.md for a beginner-friendly walkthrough (assumes no prior AI/LLM knowledge).
Mo Anisi
- OpenAI — GPT-4o, GPT-4.1, o3, o4-mini, o1, and more
- Anthropic Claude — Claude Sonnet 4, Opus 4, Claude 3.7/3.5
- Google Gemini — Gemini 2.5 Pro/Flash, 2.0, 1.5
- DeepSeek — deepseek-chat, deepseek-reasoner, deepseek-coder
- Mistral — Large, Medium, Small, Codestral, Pixtral, Ministral
- Kimi (Moonshot AI) — moonshot-v1-auto, 8k/32k/128k, kimi-latest
- Grok (xAI) — Grok 3, Grok 3 Mini, Grok 2
- Ollama — Any local model, completely free, no API key
- Custom — Any OpenAI-compatible API (LM Studio, vLLM, LocalAI, etc.)
- User-defined models — Type any model ID directly, including newly released models not yet in the built-in list
- Import RIS (PubMed/Scopus/WoS), BibTeX, and CSV reference files
- Automatic deduplication via exact DOI match + fuzzy title matching (≥90 % Levenshtein)
- Abstract-level screening before downloading PDFs (saves time & API cost)
- PDF text extraction cascade: pymupdf4llm → pdfplumber → PyPDF2
- Anti-hallucination: Pydantic schemas + instructor library + Quote-Then-Answer prompts
- Parallel processing with configurable workers
- Intelligent JSON caching to avoid reprocessing
- Multiple output formats: Excel (colour-coded), CSV, summary reports
- Define your own inclusion/exclusion screening criteria
- Specify exactly what data fields to extract
- 5 built-in domain templates (Generic, Medical, Education, Environmental, Tech)
- Advanced settings: temperature, max tokens, rate limits, chunk sizes
- AI-powered criteria enhancement — use your own API key to improve criteria wording with one click
A complete browser-based interface for the SLR tool, located in the WebApp/ directory. No desktop installation required for end users — just a running Python server.
| Feature | Description |
|---|---|
| 4-stage pipeline UI | Configure → Ingest → Process → Results — mirrors the PRISMA workflow |
| AI Enhance buttons | One-click AI improvement for screening criteria, screening prompts, and extraction fields using your configured API key |
| PDF file manager | Upload PDFs, view the list, open in browser, delete individually or clear all |
| Custom model input | Override the model dropdown with any model ID (e.g. gpt-4.5, claude-opus-4, llama3.3:70b) |
| Real-time monitor | Live progress, KPI counts, processing log streamed to the browser |
| Help & Guide drawer | 9-topic in-app guide covering Quick Start, all stages, providers, cost guide, FAQ, and a detailed disclaimer |
| Interactive disclaimer | Pre-publication checklist with interactive checkboxes |
| Auto-save settings | Provider, API key, model, and criteria persist between sessions |
| PRISMA summary | Identified → Screened → Included/Excluded counts displayed on the Results page |
Windows (one-click):
REM From the WebApp directory:
run.batManual (any OS):
# Activate your virtual environment first:
source .venv/bin/activate # macOS / Linux
.venv\Scripts\activate # Windows
cd WebApp
python app.pyThen open http://127.0.0.1:5000 in your browser.
pip install flask flask-cors
# All other dependencies are shared with the main project (requirements.txt)Windows (one-click):
REM Double-click setup.bat — creates a virtual environment and installs everything
setup.batManual (any OS):
git clone https://github.com/sadeghanisi/SLR.git
cd SLR
python -m venv .venv
# Windows:
.venv\Scripts\activate
# macOS / Linux:
source .venv/bin/activate
pip install -r requirements.txtDesktop GUI:
python slr_gui.py
# Windows shortcut: double-click launch_gui.batWeb App (browser-based):
cd WebApp
python app.py
# Then open http://127.0.0.1:5000- Select a provider → enter API key → choose a model → Test Connection
- Ollama (free): install from ollama.ai, run
ollama serve, pull a model (ollama pull llama3), select "Ollama (Local)" — no API key needed.
- Ingestion tab (GUI) or Stage 2 (Web App) → load your RIS/BIB/CSV export → Deduplicate → run abstract screening → export included records
- Setup tab (GUI) or Stage 3 (Web App) → upload PDFs → customise criteria → Start Processing
- Monitor progress live, view results, export to Excel
| Provider | Free Tier? | Recommended Model | Sign-up |
|---|---|---|---|
| OpenAI | Trial credit | gpt-4o-mini | platform.openai.com |
| Anthropic | — | claude-sonnet-4-20250514 | console.anthropic.com |
| Google Gemini | ✅ Yes | gemini-2.5-flash | aistudio.google.com |
| DeepSeek | ✅ Generous | deepseek-chat | platform.deepseek.com |
| Mistral | ✅ Free tier | mistral-large-latest | console.mistral.ai |
| Kimi (Moonshot) | ✅ Free tier | moonshot-v1-auto | platform.moonshot.cn |
| Grok (xAI) | — | grok-3-mini-fast | console.x.ai |
| Ollama | ✅ Fully free | llama3 / mistral | ollama.ai |
| Custom | Varies | — | — |
Cost estimates:
| Approach | Est. Cost | Best For |
|---|---|---|
| Ollama (local) | $0 | Large projects, privacy-sensitive data |
| DeepSeek / Gemini | Free tier | Budget-friendly cloud option |
| GPT-4o-mini | ~$0.05–0.20 / 1,000 papers | Good balance of cost & quality |
| GPT-4o / Claude Sonnet 4 | ~$0.50–5.00 / 1,000 papers | Highest quality |
Tip: Always test with 5–10 PDFs before processing your full corpus.
SLR/
├── slr_gui.py # Desktop GUI (Tkinter, 5 tabs)
├── housing_enhanced.py # Core automation engine
├── llm_interface.py # Universal LLM provider interface (9 providers)
├── ingestion.py # Reference import, dedup & abstract screening
├── prompt_editor.py # Screening/extraction criteria editor (GUI)
├── advanced_config.py # Advanced settings dialog (GUI)
├── custom_models.json # User-added model IDs (auto-created on first use)
├── test_tool.py # Basic test suite
├── requirements.txt # Python dependencies
├── setup.bat # Windows one-click setup
├── install_dependencies.bat # Alternative dependency installer
├── launch_gui.bat # Windows GUI launcher
├── COMPLETE_USER_GUIDE.md # Detailed beginner-friendly guide
├── README.md # This file
├── LICENSE # MIT License
├── docs/ # GitHub Pages website
└── WebApp/ # Browser-based Web Application (v3.3.0)
├── app.py # Flask backend (19+ API routes)
├── run.bat # Windows one-click launcher
├── requirements.txt # Web app dependencies (Flask, flask-cors)
├── templates/
│ └── index.html # Single-page application
├── static/
│ ├── css/style.css # Academic research design system
│ └── js/app.js # Frontend logic
└── uploads/ # Uploaded files (gitignored)
| Problem | Solution |
|---|---|
| "Failed to initialise LLM manager" | Check API key & internet; click Test Connection |
| "Rate limit exceeded" | Increase rate-limit delay; reduce Max Workers |
| Ollama connection failed | Run ollama serve, verify http://localhost:11434, check ollama list |
| Missing module error | Run pip install -r requirements.txt inside your venv |
| Web App: "No module named flask" | Run pip install flask flask-cors in your venv, or use run.bat |
| Provider list not loading (Web App) | Hard-refresh browser (Ctrl+Shift+R); check the server terminal for errors |
Pull requests are welcome! Key extension points:
- New LLM provider →
llm_interface.py— subclassLLMProvider - Domain template →
prompt_editor.py— add toTemplateSelector - GUI enhancement →
slr_gui.py - Web App →
WebApp/app.pyandWebApp/static/ - Extraction logic →
housing_enhanced.py
Please open an issue before major changes.
MIT — free to use, modify, and distribute with attribution.
- Read the Complete User Guide
- Check the in-app Help panel (Web App) or Help tab (GUI)
- Review logs in the output folder
- Run the test suite:
python test_tool.py - Open an issue on GitHub
IMPORTANT — Please read before use.
"As Is" Provision. This software is provided "as is," without warranty of any kind, express or implied. The authors and distributors accept no responsibility for decisions made based on AI-generated screening or extraction results.
AI Limitations. This tool assists with — but does not replace — human judgment. AI models can and do make errors, including incorrect inclusion/exclusion decisions and inaccurate data extraction. All AI outputs must be independently verified by qualified researchers before use in any publication, thesis, clinical decision, or policy document.
No Academic Guarantee. Use of this tool does not ensure compliance with PRISMA, CONSORT, or any other reporting standard. Researchers remain solely responsible for the methodological integrity, transparency, and accuracy of their systematic reviews.
Data Privacy. When using cloud-based AI providers (OpenAI, Anthropic, Google, DeepSeek, or others), your paper content is transmitted to third-party servers. The authors of this tool make no representations regarding how those providers store, process, or use your data. Consult each provider's privacy policy before processing sensitive or unpublished material. For confidential data, use the local Ollama option.
Cost and Billing. API usage fees are charged directly by third-party AI providers. The authors of this tool have no visibility into, or responsibility for, charges incurred through your API account. Monitor your usage and set billing limits with your provider before running large processing jobs.
Institutional Compliance. It is your responsibility to verify that AI-assisted research methods comply with your institution's policies, your funding body's requirements, and the ethical standards of your field. Disclose AI tool usage in all relevant sections of your research output.
No Liability. To the fullest extent permitted by law, the authors, contributors, and distributors of this tool shall not be liable for any direct, indirect, incidental, or consequential damages arising from its use, including but not limited to data loss, incorrect research conclusions, academic penalties, or financial charges.
By using this tool, you acknowledge that you have read, understood, and accepted these terms.