> tpipe/docs

Documentation Hub

Everything you need to build with TPipe — from first steps to advanced P2P orchestration.

Browse by Section Quick Start

Getting Started

2 docs

Install TPipe, configure your first Reasoning Pipe, and compose Pipes into Pipelines.

Core Concepts

25+ docs

The foundational TPipe abstractions — Reasoning Pipes, ContextBank, Pipe/Pipeline classes, KillSwitch, DITL.

Containers

8 docs

Multi-agent orchestration patterns: Manifold, Junction, Connector, Splitter, MultiConnector, DistributionGrid.

Advanced Concepts

10+ docs

P2P Registry, routing, transport, and descriptors for distributed agent communication.

API Reference

20+ docs

Complete reference for all TPipe public classes, methods, and configuration options.

Bedrock

3 docs

AWS Bedrock integration, Claude 3 and GPT-OSS inference, guardrails, IAM role configuration.

Ollama

1 doc

Run local LLM inference with Ollama — model selection, Stdio transport, local embedding support.

OpenRouter

1 doc

OpenRouter API integration, multi-model routing, unified inference across providers.

Case Studies

2 docs

Real-world TPipe deployments: TStep debugging, TPipeWriter manuscript generation, headless server use cases.

Comparison

1 doc

Feature-by-feature comparisons against adjacent technologies: Apache Camel, LangChain, LangGraph, and others.

Superpowers

1 doc

Authoritative design specs for TPipe container and execution-loop architectures. Living documents that drive implementation.

GenericOpenAI

1 doc

TPipe-GenericOpenAI pipe for any OpenAI-compatible API: OpenAI, Azure, Anthropic, DeepSeek, Groq, Together, and custom proxies.
> tpipe/categories

Browse by Category

Four unbranded cluster pages, each one targeting a distinct search intent. Pick the one that matches your problem.

Kotlin AI Agent Framework

JVM-native, headless-first, deterministic. Production-grade agent framework for Kotlin teams.

AI Agent Orchestration in Kotlin

Manifold, Junction, DistributionGrid, P2P. Multi-agent orchestration without a coordinator.

Agent Operating Environment

Substrate, not framework. Agents inhabit TPipe instead of invoking it.

Deterministic AI Agents

Same input, same pipe configuration, same output. By design.