{"repo":"christopheduc-me/ai-agent-boilerplate","free":true,"listed":false,"github":"https://github.com/christopheduc-me/ai-agent-boilerplate","clone":"git clone https://github.com/christopheduc-me/ai-agent-boilerplate.git","description":"Production-shaped boilerplate for AI agent web apps: Rust/Axum API, Python LangChain + Celery workers, Vue 3 SPA. Hexagonal architecture, TDD, JWT + refresh rotation, reliable job lifecycle with retries, rate limiting, structured logs, Docker & CI/CD","language":"Rust","stars":13,"topics":["ai","ai-agents","axum","langchain","boilerplate","celery","hexagonal-architecture","rust"],"license":"MIT","category":"ai-agents","readme_excerpt":"AI Agent Boilerplate A production-shaped, fully documented boilerplate for building AI-agent web applications — Rust API, Python agent workers, Vue frontend, wired together with the plumbing real deployments need. The demo product is deliberately simple: users sign up, enter a keyword, and launch a research that classifies and summarizes each finding and renders everything as a timeline sorted by publication date (live-updated over SSE) — in either of two modes, side by side. Workflow : the fixed pipeline (one search, enrich, sort). Agent : an agentic loop where the LLM picks its own queries, judges coverage, refines, asks you a clarification question when the goal is ambiguous, decides when to stop and reviews its own results before delivering, streaming its decision journal live to the browser. The value is everything around it — the patterns you would otherwise rebuild from scratch on every agent project. What you get - Hexagonal architecture on both server bricks — pure domains with zero infrastructure dependencies; use cases depend on ports (Rust traits / Python Protocols); adapters implement them. Swapping the LLM, the search provider, or the database is configuration, not surgery — including running the agent on a local model ( AGENT LLM BACKEND=ollama , ADR-041) instead of the Anthropic API, with an evaluation harness to compare models on the agent's own tasks (ADR-045). - TDD throughout — 190+ tests; the domain is tested with fakes of the ports, and no test ever call","default_branch":null,"files":null,"tree":[],"storefront":"/r/christopheduc-me","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/christopheduc-me/ai-agent-boilerplate/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}