Architecture
weaveIntel is built as a layered runtime. Each layer is independently configurable, testable, and replaceable — giving you complete control over your AI infrastructure.
Runtime Layers
Each layer handles a specific concern. Together, they form a complete AI runtime that scales from a single developer to enterprise deployments.
APIs
Interact with weaveIntel at the level that suits your needs — from high-level REST to low-level runtime hooks.
OpenAPI 3.1 compliant REST API for Loom management, skill invocation, policy configuration, and analytics. Authentication via API keys or JWT.
Type-safe TypeScript/JavaScript SDK with full IntelliSense support. Programmatic access to every runtime capability with async/await patterns.
Command-line interface for project scaffolding, Loom management, deployment, and monitoring. Ideal for CI/CD pipelines and automation scripts.
Principles
The architectural decisions that guide every line of weaveIntel code.
Every AI interaction passes through governance layers before execution. Compliance is not an afterthought — it’s the foundation.
No vendor lock-in. Switch between OpenAI, Anthropic, Google, or local models without changing application code. Your AI, your choice.
Every token, every decision, every cost is tracked automatically. Full distributed tracing, metrics, and audit logs out of the box.
Skills, policies, adapters, and UI components are all plugins. Extend or replace any part of the runtime without forking.
Horizontal scaling with zero configuration. From a single process to a distributed cluster, weaveIntel adapts to your workload.
Great DX matters. Type-safe APIs, hot reload, comprehensive docs, and clear error messages. Build with confidence and speed.
Dive into the source code, read the architecture docs, or scaffold your first project.