Open-source agentic API server that exposes OpenAI-compatible endpoints and routes requests to any model or backend.
schmitech/orbit
Self-hosted AI gateway that connects files, databases, vector stores and APIs to local or hosted LLMs through one OpenAI-compatible API.
Open-source alternative to:LiteLLM
Open WebUI
- Stars
- 361
- Forks
- 59
- Open issues
- 0
- Last commit
- 3 days ago
What is it?
ORBIT is a self-hosted gateway that sits between your applications and language models. It exposes one OpenAI-compatible API in front of local models (such as Ollama) and hosted providers, and adds retrieval over your own data. Adapters defined in YAML connect files, SQL and NoSQL databases, vector stores and external APIs to the same endpoint.
Beyond retrieval-augmented generation, ORBIT supports tool-calling agents. It can connect MCP tools, route requests to specialized skills, and coordinate agents over A2A or async messaging. It also covers document and image chat, voice, and structured decisions such as routing and moderation scores. An admin dashboard handles API keys, users, quotas, audit logs and cost estimates.
ORBIT is written in Python, released under Apache 2.0, and can run on-premises or in a private cloud, including air-gapped setups with local models.
Key features
- 01One OpenAI-compatible API in front of local and hosted models
- 02YAML-configured adapters for files, SQL, NoSQL, vector stores and APIs
- 03Hybrid retrieval with reranking over your own data
- 04MCP tool calling and skill routing for agents
- 05A2A and async messaging between agents
- 06Admin UI with SSO, role-based access, PII detection and audit logs
Use cases
- Chatting with internal documents without sending data to a third-party service
- Letting a chatbot answer natural-language questions over a SQL database
- Running an air-gapped AI assistant with local models
- Giving one team a shared, quota-controlled LLM endpoint with usage tracking
Quick start
- 1
Pull the Ollama flavor image (no API key needed)
docker pull schmitech/orbit-ollama:latest - 2
Start the ORBIT container
docker run -d --name orbit -p 127.0.0.1:5173:5173 -p 127.0.0.1:3000:3000 -v orbit-data:/orbit/data -v orbit-models:/orbit/models schmitech/orbit-ollama:latest - 3
Open OrbitChat at http://localhost:5173 once startup completes
Things to know
- The default Docker flavor bundles Ollama, so local models need enough hardware to run.
- The admin password is generated at first start and must be read from the container log.
In collections
Topics
FAQ
- What is orbit?
- Self-hosted AI gateway that connects files, databases, vector stores and APIs to local or hosted LLMs through one OpenAI-compatible API. ORBIT is a self-hosted gateway that sits between your applications and language models. It exposes one OpenAI-compatible API in front of local models (such as Ollama) and hosted providers, and adds retrieval over your own data. Adapters defined in YAML connect files, SQL and NoSQL databases, vector stores and external APIs to the same endpoint.
- Is orbit free to use? What license does it have?
- orbit is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is orbit still maintained?
- Actively maintained. The last commit to orbit was on Nov 4, 2026.
- What can I use orbit for?
- Chatting with internal documents without sending data to a third-party service Letting a chatbot answer natural-language questions over a SQL database Running an air-gapped AI assistant with local models Giving one team a shared, quota-controlled LLM endpoint with usage tracking
- How do I install orbit?
- Pull the Ollama flavor image (no API key needed): docker pull schmitech/orbit-ollama:latest Start the ORBIT container: docker run -d --name orbit -p 127.0.0.1:5173:5173 -p 127.0.0.1:3000:3000 -v orbit-data:/orbit/data -v orbit-models:/orbit/models schmitech/orbit-ollama:latest Open OrbitChat at http://localhost:5173 once startup completes
- Where can I get orbit?
- The source code and installation instructions are on GitHub: https://github.com/schmitech/orbit
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An experimental Go platform for long-horizon workflow automation using teams of AI assistants, with a Vue web UI.
A private desktop AI studio for macOS, Windows and Linux that runs open models locally, with an OpenAI-compatible gateway and no account or cloud.
Open-source Go gateway that routes LLM requests across 30 providers through one OpenAI-compatible API, with caching, guardrails and cost controls.
Self-hosted Go AI gateway exposing an OpenAI- and Anthropic-compatible API in front of many LLM providers, with cost tracking, failover and guardrails.
Self-hosted, OpenAI-compatible LLM gateway that routes one endpoint to 40+ providers with virtual keys, budgets, and usage tracking.