AI-powered tools for coding and content creation
TokenRing AI builds practical, agent-driven software for people who want to create faster with AI. Our ecosystem centers around:
- TokenRing One — The all-in-one AI assistant for development and content creation
- llama-scale — An OpenAI-compatible LLM router for locally hosted inference servers
TokenRing One sits on top of a modular TypeScript monorepo with a large shared package ecosystem for agents, workflows, research, storage, publishing, automation, and real-time interfaces. llama-scale is a lightweight Rust service that unifies local model backends behind a single endpoint.
A unified AI-powered platform built for developers and creators who want to work directly with their codebase and content through chat, commands, automation, and specialized agents.
What it does:
- Coding: Helps with editing, refactoring, testing, and debugging
- Writing: Assists with article writing, editing, and publishing workflows
- Research: Supports web search, Wikipedia, and information retrieval
- Infrastructure: Works locally with your code, databases, and publishing platforms
- Integration: Connects with git, Docker, Kubernetes, WordPress, Ghost, and more
Highlights:
- Multiple AI providers (OpenAI, Anthropic, Google, Groq, Cerebras, DeepSeek, and more)
- Interactive CLI and web interface
- Specialized agents for frontend, backend, testing, DevOps, architecture, security, and content
- WebSocket and remote communication support
- Strong local-first development and workflow state
- Agent-based content creation and code orchestration
A simple, Rust-based OpenAI-compatible LLM router for locally hosted inference servers. Point your apps at one endpoint and llama-scale forwards each request to Ollama, llama.cpp, vLLM, LM Studio, or any combination of them.
What it does:
- Routing: Load-balances across backend replicas with least-connections selection
- Session affinity: Keeps multi-turn conversations pinned to the same backend so context stays warm
- Failover: Health-checks backends and fails over by fallback tier when hosts are down or saturated
- Compatibility: Exposes a standard OpenAI
/v1API so existing clients work without changes - Ops: Supports Docker, Kubernetes, systemd packages, TLS, multi-user API keys, and Prometheus metrics
Highlights:
- OpenAI-compatible passthrough for any
/v1/*path (streaming supported) - Conversation-sticky routing with no client changes
- Merged
/modelsendpoint with background refresh and per-backend aliases - Per-backend concurrency caps and stream-aware timeouts
- Bearer API key auth (including multi-user model allowlists and rate limits)
- Install via script, npm/npx, prebuilt binaries, Cargo, Docker, or
.deb/.rpm
TokenRing One is powered by the broader TokenRing AI monorepo: a modular TypeScript ecosystem with 50+ packages for agent orchestration, AI clients, storage, filesystem abstraction, research, scripting, workflow execution, publishing, cloud services, and frontend interfaces.
Across the TokenRing platform, the ecosystem includes:
- Agent orchestration for specialized AI workflows
- Multi-provider AI support across leading commercial and open models
- Local LLM routing with OpenAI-compatible load balancing and session affinity
- Local and cloud filesystem support
- Research and web search tooling
- Database and persistence layers
- Publishing and communication integrations
- CLI, web UI, and API-driven interfaces
- Workflow automation and scripting
- Languages: TypeScript, JavaScript, Node.js, Bun, Rust
- Frontend: React, Vite
- Infrastructure: Docker, Kubernetes, WebSockets, Fastify, systemd
- Storage: SQLite, MySQL, PostgreSQL, S3
- AI: OpenAI, Anthropic, Google, Groq, Cerebras, DeepSeek, Ollama, llama.cpp, vLLM, LM Studio, and more
- Explore TokenRing One for the complete package ecosystem
- Explore llama-scale to route local LLM traffic through one OpenAI-compatible endpoint
We welcome contributions across our open-source projects. Check the individual repositories for setup instructions, contribution guidelines, and package-specific documentation.
- Website: tokenring.ai
- GitHub: github.com/tokenring-ai
TokenRing AI builds practical agent-driven tools for developers and content creators.