Pipelex/pipelex

Declarative language for composable Al workflows. Devtool for agents and mere humans.

What it solves

Pipelex provides a way to create reusable, typed, and repeatable AI procedures called "Methods." It solves the problem of fragile, non-deterministic AI workflows by introducing a declarative language (.mthds files) that allows developers to define structured inputs, outputs, and multi-step pipelines that can be executed across different LLM providers without rewriting the logic.

How it works

Users define AI Methods using a TOML-like syntax in .mthds files. These methods can be simple single-prompt LLM calls (PipeLLM) or complex sequences of steps (PipeSequence), including document extraction (PipeExtract). Pipelex handles the model routing (supporting 60+ models), structured output parsing, and orchestration. These methods can be executed via a CLI, Python SDK, TypeScript SDK, or as tools for AI agents via the Model Context Protocol (MCP).

Who it’s for

It is designed for developers and AI engineers who need to build production-ready AI pipelines that are model-agnostic, composable, and easy to share or version.

Highlights

  • Declarative Syntax: Uses human-readable .mthds files to define AI logic.
  • Typed Concepts: Implements semantic types to ensure inputs and outputs connect correctly across pipeline steps.
  • Model Agnostic: Supports 60+ models via the Pipelex Gateway, local AI (Ollama, vLLM), or direct API keys.
  • Composable: Allows nesting methods within other methods and chaining pipes into sequences.
  • Multi-Target Execution: Can be run via CLI, Python, Node.js, REST API, or integrated into workflow tools like n8n.

Related

  • Project
    growthxai/output

    A TypeScript framework for building AI workflows and agents that integrates prompt management, tracing, and evaluation into a single codebase to avoid SaaS fragmentation.

  • Project
    PaddlePaddle/PaddleX

    A low-code AI development tool based on PaddlePaddle that provides pre-trained model pipelines for rapid training, inference, and deployment across diverse hardware.

  • Project
    argilla-io/distilabel

    A framework for synthesizing synthetic data and AI feedback to create high-quality datasets for training and fine-tuning AI models.

  • Project
    pipecat-ai/pipecat

    Pipecat is an open‑source Python framework for building real‑time voice and multimodal AI agents. It offers plug‑and‑play support for dozens of speech‑to‑text, large‑language‑model, text‑to‑speech, and speech‑to‑speech services, composable pipelines, multi‑agent orchestration, and client SDKs for web, mobile, and embedded platforms. The Pipecat CLI can scaffold a working bot in minutes, and additional tools (Whisker debugger, Tail dashboard, Voice UI Kit) help with development and deployment.

  • Project
    generative-computing/mellea

    A Python library for building predictable AI workflows by turning typed functions into structured LLM calls with automatic validation and retries.