# Aixgo

> Aixgo is a production-grade AI agent framework for Go. It enables secure, scalable multi-agent systems with no Python dependencies, single-binary deployment under 20MB, sub-100ms cold starts, and true parallel concurrency (no GIL).

Aixgo targets backend Go engineers, DevOps teams, and data engineers who need to ship AI agents in production without the operational overhead of Python frameworks like LangChain or LlamaIndex. The framework provides 13 orchestration patterns (Supervisor, Sequential, Parallel, Router, Swarm, Hierarchical, RAG, Reflection, Ensemble, Classifier, Aggregation, Planning, MapReduce), 6 agent types (ReAct, Classifier, Aggregator, Planner, Producer, Logger), 8+ LLM providers (OpenAI, Anthropic, Gemini, xAI, Vertex AI, Amazon Bedrock, HuggingFace, plus inference services), Pydantic AI-style validation retry for structured output, and Model Context Protocol (MCP) tool calling.

Every link below is read out of this site's own pages and data files when the site builds, and a CI gate fails the build when the result stops matching what was published. The full text of every guide is at https://aixgo.dev/llms-full.txt.

## Products

- [Aixgo](https://aixgo.dev/): The framework itself. Agent runtime, supervisor, orchestration patterns and provider integrations, in one Go module.
- [Aixgo Code](https://aixgo.dev/code/): An autonomous coding agent in your own GitHub. Issue in, pull request out, a human merges. It runs in your Actions, on your runners, with your secrets. (Beta · v0.5.0 · Apache-2.0)
- [Aixgate](https://aixgo.dev/aixgate/): A deny-by-default sandbox for AI coding agents. Run Claude Code, Cursor, Aider or OpenAI Codex without giving them your .env, ~/.aws or ~/.ssh, blocked at the syscall boundary, even under prompt injection. (v0.1 proof of concept · MIT)

## Documentation

- [Quick Start Guide](https://aixgo.dev/guides/quick-start/): Get started with Aixgo in under 5 minutes. Build your first multi-agent system.
- [Core Concepts](https://aixgo.dev/guides/core-concepts/): Understand Aixgo's fundamental building blocks: agent types, supervisor patterns, and message-based communication.
- [Multi-Agent Orchestration](https://aixgo.dev/guides/multi-agent-orchestration/): Learn how to coordinate multiple agents with supervisor patterns and message routing.
- [Single Binary vs Distributed Mode](https://aixgo.dev/guides/single-vs-distributed/): Understand how Aixgo scales from local development to distributed production with zero code changes.
- [Type Safety & LLM Integration](https://aixgo.dev/guides/type-safety/): Leverage Go's type system for compile-time guarantees in LLM interactions and tool definitions.
- [Agent Types Guide](https://aixgo.dev/guides/agent-types/): Comprehensive guide to all Aixgo agent types including Classifier and Aggregator agents with examples and best practices.
- [Interactive Coding Assistant](https://aixgo.dev/guides/chat-assistant/): Use the Aixgo chat command to interact with an AI coding assistant that can read and write files, run git operations, and execute terminal commands
- [Session Persistence](https://aixgo.dev/guides/sessions/): Build AI agents that remember with built-in session persistence, checkpoints, and multiple storage backends
- [Validation with Automatic Retry](https://aixgo.dev/guides/validation-with-retry/): Pydantic AI-style automatic validation retry for 40-70% improved structured output reliability
- [Vector Databases in Aixgo](https://aixgo.dev/guides/vector-databases/): Complete guide to building RAG systems with vector databases and embeddings
- [Embeddings in Aixgo](https://aixgo.dev/guides/embeddings/): Complete guide to choosing and using embedding providers for RAG, semantic search, and similarity matching
- [Pattern Composition for Multi-Phase Workflows](https://aixgo.dev/guides/pattern-composition/): Compose existing orchestration patterns to build complex multi-phase workflows without custom orchestrators
- [Building Docker Images from Scratch](https://aixgo.dev/guides/docker-from-scratch/): Create minimal <20MB production containers leveraging Aixgo's single binary advantage.
- [Production Deployment](https://aixgo.dev/guides/production-deployment/): Deploy Aixgo agents to production with best practices for scaling and monitoring.
- [Observability & Monitoring](https://aixgo.dev/guides/observability/): Instrument Aixgo agents with OpenTelemetry, distributed tracing, and structured logging for production visibility.
- [Cost Optimization Strategies](https://aixgo.dev/guides/cost-optimization/): Optimize LLM costs using caching, monitoring, intelligent routing, and deterministic aggregation
- [Goals, gates, and loops](https://aixgo.dev/guides/goals-gates-loops/): A method for building artifacts that maintain themselves, taught through one running example: the release line on this homepage, the gate that grades it, and the four defects the first drill caught.
- [Provider Comparison](https://aixgo.dev/guides/provider-comparison/): Compare vectorstore and embedding providers to choose the right stack
- [Provider Integration Guide](https://aixgo.dev/guides/provider-integration/): Integrate Aixgo with OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock, HuggingFace, Ollama (local), and vector databases.
- [AWS Bedrock Integration Guide](https://aixgo.dev/guides/aws-bedrock/): Integrate Aixgo with Amazon Bedrock for enterprise-grade AI models with single API access, regional deployment, and AWS security.
- [Extending Aixgo](https://aixgo.dev/guides/extending-aixgo/): Guide to adding custom LLM providers, vector databases, and embedding services
- [Using Public Interfaces](https://aixgo.dev/guides/using-public-interfaces/): Build custom agents and integrate Aixgo into existing Go applications with the public agent package

## About the framework

- [Aixgo Features](https://aixgo.dev/features/): What is implemented across agents, LLM providers, security, observability and infrastructure, with the numbers that go with each.
- [Aixgo v1 Compatibility Promise](https://aixgo.dev/v1-compatibility/): Our commitment to API stability, upgrade paths, and long-term maintenance for Aixgo v1.0 and beyond.
- [Orchestration Patterns](https://aixgo.dev/features-patterns/): Production-proven agent orchestration patterns in Aixgo
- [The Aixgo Proverbs](https://aixgo.dev/proverbs/): 15 production-tested principles for building AI agents that ship and scale. Inspired by Go Proverbs and The Zen of Go.
- [Why Aixgo](https://aixgo.dev/why-aixgo/): Why the framework is written in Go, what that costs, and where a Python framework is still the better answer.

## Blog

- [Aixgo v0.7.4: the chat CLI is scriptable](https://aixgo.dev/blog/v0.7.4-chat-cli-scriptable/): Catching up on v0.7.1 through v0.7.4: one-shot chat, aixgo doctor, output caps, shell completion, and a round of security fixes.
- [Aixgo v0.7.0: Amazon Bedrock Integration](https://aixgo.dev/blog/v0-7-0-release/): Access Claude, Llama, Nova, and more via AWS with enterprise-grade security
- [Aixgo v0.6.0: Interactive Coding Assistant with Multi-Model Support](https://aixgo.dev/blog/v0-6-0-release/): Interactive coding assistant with file operations, git integration, and cost tracking
- [Aixgo v0.5.0: Public Provider API and Guided ReAct Workflows](https://aixgo.dev/blog/v0-5-0-release/): LLM providers now available as public API for external projects, plus guided execution workflows with step-by-step verification for improved ReAct agent reliability
- [Aixgo v0.4.0: Go 1.26, Advanced Planning Strategies, Enhanced Security, and RAG Variants](https://aixgo.dev/blog/v0-4-0-release/): Major release with Go 1.26 upgrade, 6 advanced planner strategies (MCTS, Tree-of-Thought, ReAct Planning), 4 RAG pattern variants, JWT verification, file-based API keys, and typed MCP tool registration
- [Aixgo v0.3.3: Session Persistence, Runtime Consolidation, and Security Hardening](https://aixgo.dev/blog/v0.3.3-session-persistence/): AI agents that remember with built-in session persistence, automated binary releases, unified runtime API, distributed runtime parity, and complete security baseline
- [v0.2.6: Dependency Updates & Release Automation](https://aixgo.dev/blog/v0.2.6-dependency-updates/): Major Google GenAI SDK update, GitHub Actions modernization, and automated binary releases
- [v0.2.5: Monorepo Consolidation](https://aixgo.dev/blog/v0.2.5-monorepo-consolidation/): Website merged into main repository for unified development experience
- [Aixgo v0.2.3 & v0.2.4: Phased Agent Startup with Dependency Ordering](https://aixgo.dev/blog/v0.2.3-and-v0.2.4-phased-startup/): Introducing dependency-aware agent startup that eliminates race conditions in multi-agent systems using topological sort and phased initialization.
- [Aixgo v0.2.2: Public Interfaces for Library-Style Integration](https://aixgo.dev/blog/v0.2.2-public-interfaces/): Introducing the public agent package: build custom agents and integrate Aixgo into existing Go applications with minimal dependencies.
- [All posts](https://aixgo.dev/blog/): The index, listing all 13 posts newest first.

## Latest release

- [v0.7.5](https://github.com/aixgo-dev/aixgo/releases/tag/v0.7.5): Released 10 September 2026. A scheduled workflow writes this fact from the GitHub API, and a gate fails when it falls behind either the latest release or the deployed page. Nobody edits it by hand.

## Source and reference

- [GitHub repository](https://github.com/aixgo-dev/aixgo): Source, issues, and the contributing guide.
- [Go package reference](https://pkg.go.dev/github.com/aixgo-dev/aixgo): Generated API documentation for the exported Go surface.
- [Releases](https://github.com/aixgo-dev/aixgo/releases): Version history and changelogs.
- [Discussions](https://github.com/orgs/aixgo-dev/discussions): Community questions and design conversations.
- [Full documentation text](https://aixgo.dev/llms-full.txt): Every guide's markdown source concatenated into one file.

## Optional

- [Roadmap](https://github.com/orgs/aixgo-dev/projects/1): Public project board and milestone tracking.
- [RSS feed](https://aixgo.dev/index.xml): New posts as they publish.