ORBIT (Open Retrieval-Based Inference Toolkit) is a middleware platform that provides a unified API for AI inference. It acts as a central gateway, allowing you to connect various local and remote AI models with your private data sources like SQL databases and vector stores.

ORBIT gives you a single, consistent API to run LLMs (local or cloud) against your private data sources with portability, performance, high-availability, and security at the core.

⭐️ If ORBIT helps you ship faster, please consider starring the repo to support the roadmap.

• ✨ Highlights
• 🚀 Quick Start
• 🛠️ Why ORBIT
• 🏗️ Architecture Overview
• ✨ What Can You Build with ORBIT?
• ⭐ Support the Project
• 📖 Documentation
• 🤝 Community & Support
• 📄 License

• Unified AI gateway that normalizes requests across local models, cloud APIs, and hybrid deployments.
• Bring-your-own data with production-grade RAG adapters for SQL, vector stores, and custom datasources.
• Secure by default with token-based auth, role-aware API keys, and pluggable content moderation.
• Ready for teams thanks to batteries-included clients (CLI, React widget, Node/Python SDKs) and automation scripts.

• Python 3.12+ (for running the server or CLI locally)
• Docker Engine 24+ (if you prefer containers)
• MongoDB (Atlas or local) to unlock authentication, RAG, and history persistence
• Optional: Redis cache plus your choice of vector DB (Chroma, Qdrant, Pinecone, Milvus)

Refer to the Docker Setup Guide or run the bundled scripts from the docker/ directory:

cd docker
chmod +x docker-init.sh orbit-docker.sh
./docker-init.sh --build --profile minimal

# Download the latest release archive
curl -L https://github.com/schmitech/orbit/releases/download/v1.5.8/orbit-1.5.8.tar.gz -o orbit-1.5.8.tar.gz
tar -xzf orbit-1.5.8.tar.gz
cd orbit-1.5.8

# Bootstrap dependencies and download a small model
cp env.example .env
./install/setup.sh --profile minimal --download-gguf granite4-micro

# Start the ORBIT server
source venv/bin/activate
./bin/orbit.sh start 

# Check the logs
cat ./logs/orbit.log

Browse to http://localhost:3000/dashboard to monitor the ORBIT server:

ORBIT Dashboard: Monitor, search, and configure your environment.

pip install schmitech-orbit-client

# Point to a running ORBIT instance (defaults to http://localhost:3000)
orbit-chat

cd clients/chat-app
npm install
npm run dev

• Create an API key tied to the adapter you want to expose (./bin/orbit.py key create).
• Enable or customize adapters in config/adapters.yaml and redeploy to connect new datasources.
• Skim the docs for deep dives on auth, configuration, and deployment patterns.

• Run securely with your data thanks to first-class support for on-prem hardware, air-gapped installs, and strict authentication defaults.
• Mix and match models (local, hosted, or API) through a single contract without rewriting downstream apps.
• Production-ready RAG with adapters for SQL, vector databases, and pipelines that keep context fresh.
• Dev-friendly tooling including a CLI, SDKs, React widget, and language clients maintained in this repo.

• Platform & infra teams who need a stable control plane for LLM workloads.
• Product teams shipping AI copilots that depend on reliable retrieval and guardrails.
• Researchers & tinkerers exploring local-first stacks or evaluating different foundation models.

Have a story or feature request? Open an issue or add it to the Roadmap.

ORBIT uses a flexible adapter architecture to connect your data to AI models. An API key is tied to a specific adapter, effectively creating a specialized "agent" for a certain task. Here are a few examples:

Provide instant, semantically-aware answers from a knowledge base. Perfect for customer support or internal documentation.

Sample Questions:

• "What are the summer camp programs for kids?"
• "How do I register for the contemporary dance class?"

NOTE: You need an instance of MongoDB to enable adapters

Here's the Sample Q/A datasets for this example. The knowledge base corresponds to a municipal services assistant.

Set inference_only mode to false in config/config.yaml:

inference_only: false

Enable the adapter in config/adapters.yaml:

- name: "qa-sql"
  enabled: true
  type: "retriever"
  datasource: "sqlite"
  adapter: "qa"
  implementation: "retrievers.implementations.qa.QASSQLRetriever"

Restart ORBIT:

./bin/orbit.sh start --delete-logs

Generate sample data and API Key (Default SQLite DB in examples/sqlite/sqlite_db):

#Login as admin first. Default password is admin123. You should change after installing ORBIT.
./bin/orbit.sh login

# Set up SQLite database with Q&A data.
./examples/sample-db-setup.sh sqlite

Start chatting with your new key:

orbit-chat --url http://localhost:3000 --api-key YOUR_API_KEY

Ask questions about your data in natural language and get answers without writing SQL.

Sample Questions:

• "Show me all users from Toronto"
• "What are the top 10 users by age?"
• "Find users created in the last month"

Install Ollama and pull the nomic-embed-text:latest embedding model. Also pull a model of choice for inference purposes.

ollama pull nomic-embed-text:latest
ollama pull gemma3:12b

Enable the contact domain sample adapter in /config/adapters.yaml:

- name: "intent-sql-sqlite-contact"
  enabled: true
  type: "retriever"
  datasource: "sqlite"
  adapter: "intent"
  implementation: "retrievers.implementations.intent.IntentSQLiteRetriever"
  inference_provider: "ollama"
  model: "gemma3:12b"
  embedding_provider: "ollama"

Start ORBIT:

./bin/orbit.sh start --delete-logs

Create an API Key for this adapter:

# Login admin credentials
./bin/orbit.sh login

# Create an API key for the SQL intent adapter
./bin/orbit.py key create \
  --intent-sql-sqlite-contact \
  --name "Contact Adapter Demo" \
  --notes "Demo using SQLite" \
  --prompt-file examples/prompts/contact-assistant-prompt.txt

# Generate sample data
python ./utils/sql-intent-template/examples/sqlite/contact/generate_contact_data.py \
  --records 500 \
  --output ./examples/sqlite/sqlite_db \
  --clean

# Test data exists
sqlite3 examples/sqlite/sqlite_db 'SELECT * FROM users LIMIT 5;'

# Start chatting with your new key
orbit-chat --url http://localhost:3000 --api-key YOUR_API_KEY

Looking for more samples? Browse the examples/ directory for data loaders, prompts, and client integrations you can adapt.

Your support keeps ORBIT independent and focused on open-source innovation.

• ⭐ Star the repo to signal that ORBIT matters to you.
• 📣 Share a demo, blog, or tweet so other builders discover it.
• 🐛 Open issues and PRs—your feedback directly shapes the roadmap.

For more detailed information, please refer to the official documentation.

• Installation Guide
• Configuration
• Authentication
• RAG & Adapters
• Development Roadmap
• Contributing Guide

• Questions? Open an issue
• Updates: Check the changelog
• Commercial Support: Contact schmitech.ai
• Maintained by: Remsy Schmilinsky

Apache 2.0 - See LICENSE for details.