> ## Documentation Index
> Fetch the complete documentation index at: https://bobbysantiago.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Writing

> Notes from the intersection of Rails and AI — and the occasional detour into developer workflow.

I write to pin down what I've actually learned, not to chase a content calendar. Two pieces so far, both things I needed and couldn't find written down clearly.

## Articles

**[How to Add AI to Your Rails App: FastAPI Integration with Local LLMs](/blog/rails-python-ai-services)**
The tutorial I wish I'd had. It walks through wiring an LLM into a Rails 8 app the way you'd actually want to in production: the call lives in a background job, with retries, timeouts, auth, and rate limiting around it, all runnable locally via Docker Compose. Every snippet is from a working repo you can clone — and it ships in "mock mode" so you can run the whole thing without a GPU or an API key.

**[tmux Workflow: From iTerm Chaos to Developer Productivity](/blog/from-iterm-chaos-to-tmux-workflow)**
An honest account of dragging my Rails dev environment out of a 12-pane, three-monitor mess and into tmux — including every mistake I made along the way. The payoff that finally sold me: detach on the desktop, `ssh` in from a laptop, reattach, and pick up mid-task with the servers still running.

## Planned

* **Streaming LLM responses in Rails** — getting tokens to the browser as they generate.
* **RAG over pgvector** — retrieval-augmented generation using Postgres as the vector store.
* **Circuit breakers for AI services** — what to do when your model provider has a bad day.