Overview
AI Engineering Toolkit (AIET) is an open-source infrastructure framework for software engineers building stateful, persistent-memory AI agents. It is local-first by default, and use of external embedding providers is explicit opt-in and may transmit data to that provider. The project addresses unstructured prompt context, context overflows, memory fragmentation, and unvetted autonomous agent mutations.
AIET provides five standardized memory primitives—Entity, Directive, Assertion, Event, and Relation—with strict Zod and JSON schemas. It includes memory candidate formation, safety and governance components, rule-based memory consolidation, a deterministic 8-stage context compiler, and hybrid RRF retrieval infrastructure. The compiler ingests, sanitizes, validates, normalizes, filters, ranks, fits, and emits context artifacts within configured token budgets, such as AGENTS.md, CLAUDE.md, and .cursorrules.
AIET is organized as a monorepo of 28 workspace projects, with 24 publishable workspace packages plus 4 runnable reference examples. It offers framework adapters and MCP integration for Vercel AI SDK, LangGraph, OpenAI Agents SDK, and Model Context Protocol for Claude Code, Cursor, and Windsurf.
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Teams use AIET to store directives and assertions and compile context for coding agents.
- Teams use AIET with LangGraph to checkpoint agent state, store findings, and detect contradictions.
- Teams use AIET with Vercel AI SDK to store customer and event data, retrieve memory, and create proposals.
- Teams use AIET with OpenAI Agents SDK to construct adapter tools, store preferences and tasks, and explain retrieval.
- Teams use AIET CLI to check system health and diagnostics.
- Teams use AIET MCP server to connect AI coding tools such as Claude Code, Cursor, and Windsurf.
- Teams use AIET to perform hybrid RRF memory search.
Features
Open source AI workflows
Reusable automation templates
Prompts for AI projects
Python engineering utilities
Developer and student resources
Reviews
No reviews yet. Be the first to share your experience!