tx
Primitives, not frameworks. Headless infrastructure for AI agents.
A small set of reusable primitives for task state, specs, memory, coordination, and observability. You keep the orchestration loop. tx gives you the building blocks.
curl -fsSL https://raw.githubusercontent.com/jamesaphoenix/tx/main/install.sh | shFive interfaces, one data model
Use whichever interface fits your workflow. They all operate on the same local SQLite database.
CLI
Shell scripts, human operators, local loops
MCP Server
Claude Code, Cursor, IDE integrations
TypeScript SDK
Custom Node/Bun agents
REST API
Language-agnostic HTTP clients
Dashboard
Visual monitoring and management
Six layers of agent infrastructure
Each layer is independent. Use one or all six. No lock-in.
Task Management
Queue, persistence, dependencies, and lifecycle tracking for agent tasks.
Spec-Driven Development
Docs-first intent and closure. Define what you want, let agents figure out how.
Memory & Context
Durable knowledge and prompt context that persists across agent sessions.
Bounded Autonomy
Controls for agent freedom: labeling, guards, and verification before execution.
Coordination
Multi-worker primitives: claims, messaging, and task distribution across agents.
Observability
Visibility via trace, health, stats, and a real-time monitoring dashboard.
Installation
Standalone Binary
Recommended. Single binary, no dependencies.
curl -fsSL https://raw.githubusercontent.com/jamesaphoenix/tx/main/install.sh | shnpm
Install via npm (requires Bun runtime).
npm install -g @jamesaphoenix/tx-cliFrom Source
Clone and build with Bun.
git clone https://github.com/jamesaphoenix/tx
cd tx && bun install && bun run buildReady to build with agent primitives?
Free and open source. Start with the docs or install in 30 seconds.