Welcome to the official documentation for ORBIT (Open Retrieval-Based Inference Toolkit) — a unified, self-hosted AI inference platform that connects your AI models to your private data sources.
GitHub • Commercial Support • Report an Issue
- ORBIT Documentation
New to ORBIT? Start here.
| Guide | Description |
|---|---|
| Quick Start | Get ORBIT running in minutes with Docker or manual installation |
| Articles & Case Studies | Deep dives into configuration and real-world use cases |
| Tutorial | Step-by-step guide to chat with your data |
| Server Setup | Detailed server configuration and deployment options |
| Configuration Guide | Complete configuration reference |
Understand how ORBIT works.
| Guide | Description |
|---|---|
| Adapters Overview | How adapters connect AI models to data sources |
| Pipeline Architecture | Request flow and processing pipeline |
| Fault Tolerance | Circuit breakers, retries, and high availability |
| Guide | Description |
|---|---|
| Adapter Configuration | Setting up and configuring adapters |
| Adapter Capabilities | Capability system for adapter behavior control |
| QA Adapters | Question-answering with SQL and vector stores |
| Intent Adapters | Natural language to SQL/API query generation |
| Composite Intent Retriever | Route queries across multiple data sources |
| Intent Agent Retriever | Function calling and tool execution capabilities |
| File Adapter | Document upload and processing |
| Passthrough Adapter | Direct conversational AI without retrieval |
Customize ORBIT for your environment.
| Guide | Description |
|---|---|
| Configuration Reference | All configuration options explained |
| Environment Variables | Required and optional environment settings |
| Inference Providers | Configure OpenAI, Anthropic, Ollama, llama.cpp, and more |
| System Prompts | Customize AI behavior with system prompts |
Connect ORBIT to your data.
| Guide | Description |
|---|---|
| SQL Databases | PostgreSQL, MySQL, SQLite, DuckDB, Oracle, SQL Server |
| MongoDB | NoSQL document database setup |
| Vector Stores | Chroma, Qdrant, Pinecone, Milvus integration |
| Vector Retriever Architecture | Technical deep dive into vector retriever implementation |
| Chroma Setup | Detailed Chroma vector database configuration |
| Elasticsearch | Full-text search integration |
| Guide | Description |
|---|---|
| REST APIs | Connect to any JSON REST API |
| GraphQL | Query GraphQL endpoints with natural language |
| Guide | Description |
|---|---|
| File Processing | PDF, DOCX, CSV, images, audio support |
| Chunking Architecture | Document chunking and embedding strategies |
| Chunking Safeguards | Safety considerations for document processing |
Audio processing and multimodal capabilities.
| Guide | Description |
|---|---|
| Audio Services | TTS, STT, transcription, and translation integration |
| Audio Client Integration | Client-side audio integration patterns |
| Whisper Integration | OpenAI Whisper for speech recognition |
| Whisper Setup | Configure Whisper for audio processing |
| Whisper Quick Reference | Quick reference for Whisper usage |
Secure your ORBIT deployment.
| Guide | Description |
|---|---|
| Authentication | User authentication and session management |
| API Keys | Create and manage API keys |
| Role-Based Access Control | Configure user roles and permissions |
For power users and contributors.
| Guide | Description |
|---|---|
| llama.cpp Server | Run GGUF models locally |
| Shimmy Server | Lightweight OpenAI-compatible inference |
| Ollama Integration | Using Ollama for local inference |
| Guide | Description |
|---|---|
| Fault Tolerance Architecture | System resilience overview |
| Circuit Breaker Patterns | Failure handling patterns |
| Troubleshooting | Debug common issues |
| Performance Tuning | Optimize for production workloads |
| Memory Management | Prevent memory leaks |
| Rate Limiting | Rate limiting and throttling architecture |
| Datasource Pooling | Connection pooling for data sources |
| Guide | Description |
|---|---|
| MCP Protocol | Model Context Protocol integration |
| Reranker Architecture | Document reranking for improved accuracy |
| Language Detection | Multilingual support |
| Conversation Threading | Sub-conversations and cached dataset reuse |
| Autocomplete | Query suggestions from intent templates |
| Conversation History | Chat history management |
| Request Context Propagation | Context handling across the pipeline |
Contribute to ORBIT or extend its capabilities.
| Resource | Description |
|---|---|
| Contributing Guide | How to contribute to ORBIT |
| Testing PRs Locally | Guide for testing pull requests |
| SQLite Schema | Database schema reference |
| Code of Conduct | Community guidelines |
| Changelog | Release history and updates |
Need help getting started?
- Step-by-Step Tutorial — Learn how to connect your data.
- GitHub Discussions — Get help from the community.
- Commercial Support — Enterprise-grade assistance from Schmitech.
- Managed Hosting — Fully managed deployments with SLA guarantees
- Custom Development — Custom adapters, integrations, and model tuning
- Enterprise Integration — Connect to your databases, APIs, and SSO
- Installation & Setup — On-premise and cloud deployment assistance
- Training & Workshops — Hands-on training for your team
- Dedicated Support — Priority response and dedicated support engineer
👉 Contact Schmitech to discuss your requirements.
ORBIT is open source under the Apache 2.0 License.
Built with ❤️ by Schmitech