Arachne is an AI runtime, developer toolchain, and open spec for defining portable agents and knowledge bases that run anywhere.
- Runtime — A multi-tenant AI gateway that proxies requests to any LLM provider (OpenAI, Azure, Ollama, and more) with built-in observability, conversation memory, and RAG inference
- Developer tools — A CLI (
arachne) for defining Agent and KnowledgeBase artifacts as YAML specs, weaving them into signed bundles, pushing to a registry, and deploying to any tenant - Portable spec — Agents and knowledge bases are defined as versioned, content-addressed artifacts with a stable schema — write once, deploy anywhere
- AI-powered products with multi-tenant isolation, per-tenant agent configuration, and encrypted trace storage
- Portable agents defined as YAML specs with system prompts, tools, and RAG knowledge — versioned and deployable via the registry
- Knowledge pipelines that chunk documents, generate embeddings, and serve vector search through a consistent RAG inference interface
| Capability | What it does |
|---|---|
| LLM Gateway | OpenAI-compatible proxy routing to OpenAI, Azure OpenAI, or Ollama |
| Agent Registry | Push/pull versioned agent and knowledge base artifacts |
| RAG Inference | Automatic embedding, retrieval, and citation injection |
| Conversation Memory | Persistent multi-turn context with automatic summarization |
| Observability | Encrypted trace recording, token analytics, latency metrics (p95/p99) |
| Multi-tenancy | Strict data isolation, subtenant hierarchy, per-tenant provider config |
| Portal | Tenant-facing web UI for managing agents, keys, and analytics |
See Running Locally — Docker Compose (recommended) or Node.js dev mode.
The arachne CLI lets you define and deploy agent artifacts from your terminal:
arachne weave agent.yaml # bundle an agent spec
arachne push myorg/my-agent:1.0 # push to registry
arachne deploy myorg/my-agent:1.0 # deploy to a tenant→ CLI Reference · How the artifact model works
| Doc | What's in it |
|---|---|
| Running Locally | Setup with Docker Compose or Node.js |
| CLI Reference | All arachne commands with examples |
| CLI Overview | Artifact model, weave pipeline, registry, deployment |
| RAG Inference | How retrieval and embedding work |
| Registry API | HTTP API for the artifact registry |
| Portal Guide | Using the tenant web portal |
| System Embedder | Configuring a default embedding provider |
| Architecture | Component design, data flows, encryption |
| Developer Guide | Source layout, database schema, API extensions |